THE AUTHORITY ON
EMERGING TECHNOLOGIES
FOR DESIGN SOLUTIONS
07.11.13 electronicdesign.com
Survey Says Design
Engineers Have Power
On Their Minds
P. 20
NEW
PRODUCTS
ADDED DAILY
$10.00 A PENTON PUBLICATION
Periodicals Postage Paid • USPS 100 Approved Poly
DIGIKEY.COM/NEW
The best inductor
selection tools.
coilcraft.com/tools
Now in a handy
pocket size.
coilcraft.com/mobile
®
WWW.COILCRAFT.COM
gy Efficiency & Technology | Defense Electronics | Motion System Design |
gn Europe | Electronic Design China | Engineering TV | Fluid Power Expo | M
yping | One Powerful Day | Combating Conterfeit Conference | Electronic D
waves & RF | Medical Design | Source ESB | Hydraulics & Pneumatics | Glob
ource | Power Electronics Technology | Mobile Development and Design | E
Defense Electronics | Motion System Design | Auto Electronics | Electronic D
n China | Engineering TV | Fluid Power Expo | Medical Silicone | Medical Pr
Combating Conterfeit Conference | Electronic Design | Machine Design | M
sign | Source ESB | Hydraulics & Pneumatics | Global Purchasing | Distribu
s Technology | Mobile Development and Design | Energy Efficiency & Techn
on System Design | Auto Electronics
|ARE
Electronic Design Europe | Electroni
WE
Fluid Power Expo | Medical Silicone | Medical Prototyping | One Powerful D
ence | Electronic Design | Machine Design | Microwaves & RF | Medical Des
umatics | Global Purchasing | Distribution Resource | Power Electronics Te
d Design | Energy Efficiency & Technology | Defense Electronics | Motion S
| Electronic Design Europe | Electronic Design China | Engineering TV | Flu
| Medical Prototyping | One Powerful Day | Combating Conterfeit Confere
Design | Microwaves & RF | Medical Design | Source ESB | Hydraulics & Pn
bution Resource | Power Electronics Technology | Mobile Development and
nology | Defense Electronics | Motion System Design | Auto Electronics | E
c Design China | Engineering TV | Fluid Power Expo | Medical Silicone | Me
Day | Combating Conterfeit Conference | Electronic Design | Machine Desig
esign | Source ESB | Hydraulics & Pneumatics | Global Purchasing | Distrib
s Technology | Mobile Development and Design | Energy Efficiency & Techn
on System Design | Auto Electronics | Electronic Design Europe | Electroni
Fluid Power Expo | Medical Silicone | Medical Prototyping | One Powerful D
ence | Electronic Design | Machine Design | Microwaves & RF | Medical Des
umatics | Global Purchasing | Distribution Resource | Power Electronics Te
d Design | Energy Efficiency & Technology | Defense Electronics | Motion S
| Electronic Design Europe | Electronic Design China | Engineering TV | Flu
WE
SUPPORT
THE DESIGN
PROCESS
| Medical
Prototyping
| One Powerful
Day | Combating
Conterfeit Confere
Design | Microwaves
& RF |INTENT
Medical Design
ESB | Hydraulics & Pn
FROM
TO| Source
ACTION
bution Resource | Power Electronics Technology | Mobile Development and
nology | Defense Electronics | Motion System Design | Auto Electronics | E
c Design China | Engineering TV | Fluid Power Expo | Medical Silicone | Me
Day | Combating Conterfeit Conference | Electronic Design | Machine Desig
Sense-ability.
Newark element14 makes it easy to find all of your sensors and transducers – fast. newark.com
DirectFET®
w
Po
er
PQFN
n
De
y
sit
D-Pak
SO-8
Select The Best MOSFET Pair to
Meet Your Power Density Needs
Scalable Solutions for DC-DC Buck Converters
The IR Advantage
Package
RDS(on) Max.
VGS=10V
(m)
QG
VGS=4.5V
(nC)
Control
High Current PQFN 5 x 6
8.1
7.8
30
Sync
High Current PQFN 5 x 6
2.1
29
IRFHM831
30
Control
PQFN 3x3
7.8
7.3
IRFHM830
30
Sync
PQFN 3x3
3.8
15
UÊ-V>>LiÊÃÕÌÃÊvÀÊ`vviÀiÌÊ«ÜiÀ
Ê `iÃÌÞÊÀiμÕÀiiÌÃ
IRFHM830D
30
Sync
PQFN 3x3
4.3
13
UÊiV>ÀÊ«iÀvÀ>Vi
Control
8.3
30
UÊ7`iÊÀ>}iÊvÊ«>V>}iÃÊ
Sync
Half-Bridge Asymmetric
PQFN 5x6
8.6
IRFH7911
3.0
34
BVDSS
(V)
Function
IRFH5306
30
IRFH5302
Part Number
IRLR8729
30
Control
D-Pak
8.9
10
IRLR8726
30
Sync
D-Pak
5.8
15
IRF8714
30
Control
SO-8
8.7
8.1
IRF8734
30
Sync
SO-8
3.5
20
IRF8513
30
Control
Half-Bridge Asymmetric
SO-8
15.5
5.7
12.7
7.6
Sync
UÊ"«Ìâi`Ê-ÞVÊ>`Ê
ÌÀÊ"-/ÃÊvÀ
Ê }ÊÃÞÃÌiÊivwÊViVÞ
UÊ}Ê}Ì>`ÊivwÊViVÞ
UÊ«ÀÛi`ÊÌiÀ>Ê«iÀvÀ>ViÊÜÌ
Ê ÀiVÌ/® ÊÌ«Ã`iÊV}
OICE
Your FIRST CH
e
for Performanc
for more information call 1.800.981.8699 or visit us at www.irf.com
Distributors :
®
THE POWER MANAGEMENT LEADER
07.11.13
VOLUME 61, ISSUE 8
InThisIssue
FEATURES
20
SURVEY SAYS DESIGN
ENGINEERS HAVE POWER
ON THEIR MINDS
A supercapacitor, dc-dc converters,
and a charging coil garner four of
the top 10 spots on our annual list
of the Top 101 Components.
28
20
WI-FI AND BLUETOOTH
RULE THE AIRWAVES
These two core wireless technologies
continue to dominate short-range
applications with new features and
benefits.
36
ADVANCED PACKAGING
DELIVERS CAPACITY AND
PERFORMANCE
High-density BGAs and 3D
stacking increase compute
and storage density.
41
MATTERFORM FOUNDERS
DISCUSS THEIR 3D SCANNER
NEWS & ANALYSIS
16
PLAYSTATION 4
LOOK MORE ALIKE
19
MERGER PROPELS
WIRED HOME
NETWORKING
If 2D printers have 2D scanners,
why shouldn’t 3D printers have 3D
scanners?
41
XBOX ONE AND
28
COLUMNS
13
EDITORIAL
Passive Components Still
See Ink And Innovation
14
SIMPLY THE BITS
THAT MATTER
Compress Climate And Physics
Datasets For Petascale Computing
EDITORIAL MISSION:
To provide the most current, accurate, and in-depth technical
coverage of the key emerging technologies that engineers need
to design tomorrow’s products today.
64
LAB BENCH
Are You Writing Safe
And Secure Software?
36
ELECTRONIC DESIGN (ISSN 0013-4872) is published monthly with an extra issue in June and October by Penton Media Inc., 9800 Metcalf Ave.,
Overland Park, KS 66212-2216. Paid rates for a one-year subscription are as follows: $120 U.S., $180 Canada, $240 International. Periodicals
postage paid at Shawnee Mission, KS, and additional mailing offices. Editorial and advertising addresses: ELECTRONIC DESIGN, 1166 Avenue of
the Americas, New York, NY 10036. Telephone (212) 204-4200. Printed in U.S.A. Title registered in U.S. Patent Office. Copyright ©2013 by Penton
Media Inc. All rights reserved. The contents of this publication may not be reproduced in whole or in part without the consent of the copyright
owner. For subscriber services or to order single copies, write to Electronic Design, PO Box 2100, Skokie, IL 60076. POSTMASTER: Send change
of address to Electronic Design, PO Box 2100, Skokie, IL 60076. Canadian Post Publications Mail agreement No. 40612608. Canada return
address: IMEX Global Solutions, P.O. Box 25542, London, ON N6C 6B2.
GO TO ELECTRONICDESIGN.COM
5
TV\ZLYJVT
;OL5L^LZ[7YVK\J[ZMVY@V\Y5L^LZ[+LZPNUZ
4VYL5L^7YVK\J[Z
4VYL5L^;LJOUVSVNPLZ
4VYL(KKLK,]LY`+H`
(\[OVYPaLKKPZ[YPI\[VYVMZLTPJVUK\J[VYZ
HUKLSLJ[YVUPJJVTWVULU[ZMVYKLZPNULUNPULLYZ
4V\ZLYHUK4V\ZLY,SLJ[YVUPJZHYLYLNPZ[LYLK[YHKLTHYRZVM4V\ZLY,SLJ[YVUPJZ0UJ6[OLYWYVK\J[ZSVNVZHUKJVTWHU`UHTLZTLU[PVULKOLYLPUTH`IL[YHKLTHYRZVM[OLPYYLZWLJ[P]LV^ULYZ
InThisIssue
07.11.13
VOLUME 61, ISSUE 8
DESIGN SOLUTION
43
TRANSACTION-BASED VERIFICATION
AND EMULATION COMBINE FOR
MULTI-MEGAHERTZ VERIFICATION
PERFORMANCE
Rx
IF
TBV is becoming the emulation mode of choice
because of its unique ability to keep pace with
VS
VF
growing verification support requirements.
DISTRIBUTION RESOURCE
51
48
SLOW RECOVERY
STILL THE NORM
Executives adjust to slower
industry conditions as they
look to new regions and
expanding markets for growth.
51
51
GLOBALIZATION,
PRODUCT OBSOLESCENCE TOP SUPPLY
CHAIN CHALLENGES
IDEAS FOR DESIGN
Distributors respond with tools
48
Graphical, Numerical
to help engineers and procure-
Techniques Size LED
ment professionals meet the
Array’s Drive
increasingly complex electron-
PRODUCT FEATURES
CAPACITOR SUPPLIERS
59
MEET GROWING
MARKET DEMANDS
Design Kit Optimizes Any SoC
60
Maximizes Smart-Phone
2.5% over the next five years,
Antenna Efficiency
61
Fast Precision SAR ADCs
Battle For Sockets
complex electronics and innovative component solutions.
56
Digital RF MEMS Capacitor
The capacitor market will grow
driven by demand for more
56
High-Performance Computing
63
Ad Index
Antenna frequency
tuning performance
50%
FEW COMPANIES
Antenna efficiency
54
ics supply chain.
READY FOR CONFLICT
MINERALS REGS
Traditional antenna
performance
40%
30%
20%
Instantaneous
performance
10%
An IHS poll reveals that more
LTE
band
GSM
band
0%
600
than a third of electronics
700 800
900
Frequency (MHz)
1000
industry firms have made no
60
plans to meet new reporting
reqirements.
Permission is granted to users registered with the Copyright Clearance Center Inc. (CCC) to photocopy any article, with the exception of those for
which separate copyright ownership is indicated on the first page of the article, provided that a base fee of $2 per copy of the article plus $1.00
per page is paid directly to the CCC, 222 Rosewood Drive, Danvers, MA 01923 (Code No. 0013-4872/94 $2.00 + $1.00). Copying done for
other than personal or internal reference use without the express permission of Penton Media, Inc. is prohibited. Requests for special permission
or bulk orders should be addressed to the editor. To purchase copies on microfilm, please contact National Archive Publishing Company (NAPC)
at 732-302-6500 or 800-420-NAPC (6272) x6578 for further information.
GO TO ELECTRONICDESIGN.COM
2008 Winner
GOLD
EXCELLENCE
IN MAGAZINE
DESIGN
2008 Winner
SILVER
EDITORIAL
AWARD OF
EXCELLENCE
7
on
TESLA LAUNCHES NATIONWIDE
SUPERCHARGER NETWORK
Tesla Motors will expand its network of Superchargers, with 27 sations expected to open through the
summer of 2013 and 98% of the U.S. and Canada
covered by 2015.
INTERVIEW:JANE
PREYTARGETSTHE
GENDER GAP IN
COMPUTING AND
ENGINEERING
ELECTRONIC PAPER PROVIDES
DESIGN FREEDOM
Also known as e-paper, electrophoretic displays
are common in e-readers today but can be used for
much more.
blogs
JOE DESPOSITO • Engineering An IndyCar Victory
EDITOR-IN-CHIEF
DON TUITE
ANALOG/POWER
Is More About Data And Simulation Than The Car’s Electronics
• SID 2013: Quantum Dots, Yes;
ASK THE EXPERT
Join Contributing Technical Expert Dave Van Ess
in our forums as he responds to your questions with
hands-on, practical advice.
Pico Projectors, Hardly Any
LOUIS FRENZEL
COMMUNICATIONS
• CTIA 2013: Your Mobile Future
Revealed
BILL WONG
EMBEDDED/SYSTEMS/
SOFTWARE
• Arm Micro Consolidation: Silicon
Labs Acquires Energy Micro
join us online
8
07.11.13 ELECTRONIC DESIGN
TOP GUNS.
IP&E SOLUTIONS FROM AVNET
Your top flight semiconductor cannot work alone.
You need world-class interconnect, passive and electromechanical components to support
your design. Avnet’s legacy in the IP&E business dates back to 1921 and is stronger than
ever today. As a worldwide distributor of IP&E products, we source from the leading global
manufacturers – and you can count on us to provide top quality components to meet your
needs. We are ready to help your products take flight.
What can we do for you? www.avnetexpress.com
Accelerating Your Success!™
1 800 332 8638 | www.avnetexpress.com |
@avnetdesignwire
MDK-ARM™ is the complete
development environment for
ARM® Cortex™-M series
microcontrollers.
EDITORIAL
EDITOR-IN-CHIEF: JOE DESPOSITO joe.desposito@penton.com
MANAGING EDITOR: RICHARD GAWEL richard.gawel@penton.com
CREATIVE DIRECTOR: DIMITRIOS BASTAS dimitrios.bastas@penton.com
ANALOG/POWER: DON TUITE don.tuite@penton.com
COMMUNICATIONS: LOUIS E. FRENZEL lou.frenzel@penton.com
DISTRIBUTION: VICTORIA FRAZA KICKHAM SourceESBeditor@penton.com
EMBEDDED/SYSTEMS/SOFTWARE: WILLIAM WONG bill.wong@penton.com
07.11.13
ART DEPARTMENT
GROUP DESIGN DIRECTOR: ANTHONY VITOLO tony.vitolo@penton.com
SENIOR ARTIST: JAMES M. MILLER jim.miller@penton.com
INTERN: COREY BAUMANN
PRODUCTION
GROUP PRODUCTION MANAGER: JUSTIN MARCINIAK justin.marciniak@penton.com
PRODUCTION MANAGER: JULIE GILPIN julie.gilpin@penton.com
AUDIENCE MARKETING
AUDIENCE DEVELOPMENT DIRECTOR: BRENDA ROODE brenda.roode@penton.com
ONLINE MARKETING SPECIALIST: RYAN MALEC ryan.malec@penton.com
ELECTRONIC DESIGN EUROPE
EDITOR: PAUL WHYTOCK paul.whytock@penton.com
ASSOCIATE EDITOR: SALLY WARD-FOXTON sally.ward-foxton@penton.com
LIST RENTALS & CIRCULATION CUSTOMER SERVICE (LIVE)
LIST RENTALS: MERIT DIRECT
FREE SUBSCRIPTION/STATUS OF SUBSCRIPTION/ADDRESS CHANGE/MISSING BACK ISSUES
T | 866.505.7173 F | 847.763.9673 electronicdesign@halldata.com
DS-5 Professional is a full featured
development solution for all
ARM Powered® platforms.
™
SALES & MARKETING
BRAND DIRECTOR, e|DESIGN: TRACY SMITH
T | 913.967.1324 F | 913.514.6881 tracy.smith@penton.com
REGIONAL SALES REPRESENTATIVES
NORTHEAST/EASTERN CANADA: DAVE MADONIA
T | 212.204.4331 F | 913.514.3966
dave.madonia@penton.com
SOUTH: BILL YARBOROUGH T | 713.636.3809 F | 713.380.5318 bill.yarborough@penton.com
NORTHWEST/NORTHERN CALIFORNIA/WESTERN CANADA: JAMIE ALLEN T | 415.608.1959 F | 913.514.3667
jamie.allen@penton.com
MIDWEST/MID-ATLANTIC: STEPHANIE CAMPANA
T | 312.840.8437 F | 913.514.3645
stephanie.campana@penton.com
PLEASE SEND INSERTION ORDERS TO: orders@penton.com
PENTON REPRINTS: WRIGHT’S MEDIA T | 877.652.5295 penton@wrightsmedia.com
CIRCULATION: CUSTOMER SERVICE T | 866.505.7173 F | 847.763.9673 electronicdesign@halldata.com
INTERNATIONAL SALES
EUROPE: MARK DURHAM T | +44 (0)7958 564137 mark.durham@penton.com
JAPAN: HIROKAZU MORITA T | +81 3 3261 4591 F | +81 3 3261 6126
TAIWAN: CHARLES LIU T | 886 2-2727 7799 F | 886 2 2728-3686
ONLINE
DIRECTOR OF DIGITAL CONTENT: PETRA ANDRE petra.andre@penton.com
DESIGN ENGINEERING & SOURCING GROUP
VICE PRESIDENT & MARKET LEADER: BILL BAUMANN
GROUP DIRECTOR OF EDITORIAL CONTENT: NANCY FRIEDRICH
GROUP DIRECTOR OF OPERATIONS: CHRISTINA CAVANO
GROUP DIRECTOR OF MARKETING: JANE COOPER
RESEARCH MANAGER: JULIE RITCHIE
MARKETING & EVENTS SPECIALIST: ADRIAN PIAZZA
MARKETING COMMUNICATIONS SPECIALIST: CYNTHIA RODRIGUEZ
PENTON MEDIA INC.
CHIEF EXECUTIVE OFFICER: DAVID KIESELSTEIN david.kieselstein@penton.com
CHIEF FINANCIAL OFFICER: NICOLA ALLAIS nicola.allais@penton.com
SENIOR VP, DESIGN ENGINEERING GROUP: BOB MACARTHUR bob.macarthur@penton.com
+1 800 348 8051
1166 AVENUE OF THE AMERICAS, 10TH FLOOR
NEW YORK, NY 10036 T | 212.204.4200
Electronic Design | Machine Design | Microwaves & RF | Medical Design | Source ESB | Hydraulics & Pneumatics | Global Purchasing |
Distribution Resource | Power Electronic | Mobile Dev & Design | Defense Electronics | Auto Electronics | Electronic Design Europe | Engineering TV
© ARM Ltd. AD364 | 01.13
07.11.13 ELECTRONIC DESIGN
Economy oscilloscopes redefined.
InfiniiVision 2000 X-Series
70 to 200 MHz
New memory depth up to 1 Mpts
New serial decode options*
New 5-year warranty
Starting at $1,258**
InfiniiVision 4000 X-Series
InfiniiVision 3000 X-Series
200 MHz to 1.5 GHz
12.1-inch touch screen
InfiniiScan Zone touch triggering
Update rate to 1 million wfms/s
Memory depth up to 4 Mpts
Starting at $5,611**
100 MHz to 1 GHz
8.5-inch display
Update rate to 1 million wfms/s
Memory depth up to 4 Mpts
Starting at $2,929**
Oscilloscopes Redefined. With its brilliant 8.5-inch screen, fast update rate, full
upgradability and 5-in-1 instrument integration, the Agilent 2000 X-Series oscilloscope
is already first in class. Now with more memory depth, new serial decode options and
5-year warranty, it’s accelerating ahead. Experience the complete InfiniiVision line for
oscilloscope features and performance you need at prices that will surprise you.
Agilent and our
Distributor Network
Right Instrument.
Right Expertise.
Delivered Right Now.
© Agilent Technologies, Inc. 2013
800-433-5700
www.alliedelec.com/agilent
See the InfiniiVision
X-Series oscilloscopes
www.alliedelec.com/lp/agilentscopes/
*I2C, SPI, CAN, LIN, RS232, UART **Prices are in USD and are subject to change
Editorial
JOE DESPOSITO
Editor-in-Chief
joe.desposito@penton.com
Passive Components Still
See Ink And Innovation
W
JOIN US ONLINE
Become a fan at
facebook.com/
ElectronicDesign.
hen people first learn about electronics, they’re usually taught how resistors,
capacitors, and inductors interact in
a circuit. Since these components have been around a
long time, you might think there’s nothing more to say
about them and certainly not much innovation going
on. Nothing could be further from the truth. Just look at our annual list of the Top
101 Components (p. 20) to see that fully 10% of them are passives.
We’ve written about these specific products and more over the past year. For
instance, in our new online Ask The Expert series, Tamara Schmitz, an applications
manager at Intersil, wrote about selecting components for prototypes. In particular,
she wrote about selecting resistors. In the photo that accompanies her article, the
resistors look like the common ones we’re all used to seeing.
But Tamara also writes about a dizzying array of choices you might find at a distributor’s Web site. She notes that “they are categorized by the way they are mounted: surface mount, through hole, chassis mount, array… or two others: precision or
accessories. Many designers just start clicking away and hope for pictures that can
be deciphered.” You can find Tamara’s full post by clicking on the Forums tab at
electronicdesign.com.
Bill Laumeister of Maxim Integrated is also writing about passives, saying maybe
they aren’t so passive after all. “These apparently passive components can, and do,
change the signal in unexpected ways since they contain parasitic portions,” he said.
As an example of the continuing innovation in passive components, I’ll point to
two resistors that were just introduced by Ohmite Manufacturing. These are the
RW5 (5 W) and RW7 (7 W) surface-mount wirewound resistors (see the figure). As
you might guess, their unique finned body style maximizes surface area and therefore cooling capacity and makes them a great fit for high-power-density applications. You can find a full description at electronicdesign.com.
GO TO ELECTRONICDESIGN.COM
Ohmite Manufacturing’s
RW5 and RW7 surfacemount wirewound resistors sport a finned body
style that makes them a
great fit for high-powerdensity applications.
13
Simply the Bits that Matter
AL WEGENER | Contributing Technical Expert
awegener@samplify.com
Compress Climate
And Physics Datasets
For Petascale Computing
M
ulticore computing results are increasingly determined by the rate at which operands are fed to
the cores and decreasingly by the speed or the
number of cores. SDRAM and flash memory bottlenecks often
slow down computations, but disk drive speeds can also determine the rate at which supercomputers deliver results.
Indeed, in high-performance computing (HPC) applications,
“time to results” can be directly related to the rate at which
numerical datasets, such as climate measurements, are read
from disk, as well as the rate at which simulation results, such as
3D physics arrays, are stored to disk drives.
TITANIC COMPUTING
mate change. Climate scientists typically begin a simulation
at a historical point where atmospheric, oceanic, and land
temperatures were once measured. These simulations combine
multiple abstracted, mathematical 3D models of the Earth’s air,
ocean, land, and ice conditions and advance the simulation one
timestep at a time.
Depending on the purpose of the climate simulation,
timestep sizes may range from minutes (weather forecasts having a daily or weekly duration) to months (climate simulations
having 100-year durations). The output of every Nth timestep
is saved for post-simulation analysis and visualization, where N
may range from 1 to 100.
The beautiful, complex color weather maps we see on television and on the Internet are the outcome of such climate simulations, as are the decadal distribution and density maps of carbon dioxide, methane, water, clouds, and ice that are consistent
with global warming.
Scientists are improving climate models in two aspects:
improved grid density and more complex boundary interactions. Since atmospheric and oceanic grid densities have both
spatial (x-y-z) and temporal (t) dimensions, improved grid density directly affects the size of intermediate timestep results.
A twofold increase per dimension in 3D spatial grid density
creates intermediate datasets that are eight times larger, exacerbating the I/O challenges of climate simulations. Similarly,
improving the accuracy of boundary interactions (such as
between the air and ocean, or between ice and water) often
Today’s supercomputers have immense computing power.
Supercomputers are ranked according to a “flops” metric that’s
calculated using the rather dated Linpack benchmark.1 Every
six months, www.top500.org updates the list of the most powerful supercomputers. In November 2012, the fastest supercomputer in the world was the Titan supercomputer at Oak Ridge
National Labs in Tennessee, clocking in at 17.6 Pflops (17.6 =
1018 floating-point operations per second).
Titan uses Intel x86 processors and Nvidia GPUs to achieve
this prodigious calculation rate. Unfortunately, a supercomputer’s Linpack rating rarely predicts the performance of everyday “codes,” as the supercomputing programmers and users call
their application-specific software.
Two application codes that require long calculation times
and generate and consume multi-petabyte
(1018 bytes) datasets are climate and physAPAX NUMERICAL ENCODING RESULTS
ics simulations. It’s not uncommon for
APAX compression ratio
climate and physics simulations to run for
Application “Codes”
Datasets tested
at “same results”
many months using hundreds of thousands of cores and to consume terabytes
Climate
Various
127 climate variables
4:1
of RAM and petabytes of disk storage. The
Sedov blast wave radius, density,
supercomputing community is actively
Physics
LULESH
4.5:1
thickness (three variables)
evaluating various techniques, including
Mixing layer thickness, growth rate,
compression, to reduce RAM, flash, and
MIRANDA
5.4:1
energy density (six variables)
disk bottlenecks in supercomputing.
Climate simulations have been in the
Laser intensity
pF3D
(incoming and back-scattered),
6.2:1
news since the late 1990s because they
plasma density (six variables)
improve society’s understanding of cli-
14
07.11.13 ELECTRONIC DESIGN
requires higher grid densities at the boundaries. More complex
models result in longer climate simulation times and more data
to be captured and visualized.
IN THE LABS
disk I/O bottlenecks by factors of four and six. The table summarizes the results of compressing climate4 and physics5 datasets using the APAX universal numerical encoder. These two
papers are the first to document that compression’s effects maintain simulation results for large, complex HPC calculations
while significantly reducing memory and disk bottlenecks.
One of Samplify’s climate simulation collaborators participated in the latest Coupled Model Intercomparison Project
Phase 5 (CMIP5) long-term climate simulation.2 This group
References
struggled to analyze the climate results it generated for CMIP5,
1. Frequently Asked Questions on the Linpack Benchmark
which were 10 times larger than its CMIP4 datasets five years
and Top500, www.netlib.org/utk/people/JackDongarra/
earlier. This group’s problem occurred not in the generation of
faq-linpack.html
climate simulation data, but rather in the post-simulation data
2. CMIP – Coupled Model Intercomparison Project Phase
analysis and visualization phase, where slow disk reads threat5 – Overview, http://cmip-pcmdi.llnl.gov/cmip5/
ened the group’s timely delivery of well-studied CMIP5 results.
3. The National Ignition Facility: Ushering in a New Age for
Lawrence Livermore National Labs (LLNL) operates a special Science, https://lasers.llnl.gov/about/nif/
research building called the National Ignition Facility3 (NIF)
4. International Supercomputing Conference Presentation
that evaluates next-generation fusion energy. At NIF, the world’s Details, www.isc-events.com/isc13_ap/presentationdetails.
most powerful laser (actually a complex of 192 high-energy
php?t=contribution&o=2217&a=select&ra=speakerdetails
lasers) impinges on a tiny, gold-plated cylinder the size of a
5. submitted to http://sc12.supercomputing.org/
pencil eraser called a “hohlraum” (the German word for “hollow space”) in which the fusion reaction occurs.
AL WEGENER, CTO and founder of Samplify, earned an MSCS
Because of the immense amounts of power and months of
from Stanford University and a BSEE from Bucknell University.
preparation required for each NIF experiment, it is significantly less expensive
for LLNL to perform multiple hohlraum
physics simulations in a supercomputer
than it is to perform a single NIF experiment. LLNL’s dedicated NIF physicists
perform considerably more NIF simulaRELIABLE CAPACITORS FOR AUTOMOTIVE APPLICATIONS
tions than experiments, but the complexity of laser-plasma interactions within
the hohlraum requires the interaction of
several physics simulations.
As with climate simulations, grid density and timestep size are key NIF simulation parameters, and LLNL physicists
always want smaller grid and timestep
spacing because better resolution provides additional physics insights. Simulating NIF’s multi-physics interactions
AEC
• Extremely low failure rate
Q200
generates petabyte-scale datasets over
QUALIFIED*
• Body and safety system grades
months-long simulations using LLNL’s
• Built with our own raw materials and advanced technology
Sequoia, the world’s second-fastest
• Quality controlled through entire manufacturing process
supercomputer.
• Smaller cutting-edge solutions at competitive prices
Samplify collaborated both with
• TS-16949 Certified Factory
Deutsches Klimarechenzentrum (DKRZ;
German Climate Computing Centre) and
*Please contact our sales
LLNL in experiments that demonstrate
office for latest information
that lossy compression of climate and
t-yuden.com/automotive
physics datasets maintains the simulation
results while reducing the memory and
800-348-2496
GO TO ELECTRONICDESIGN.COM
& Analysis
Xbox One And PlayStation 4
Look More Alike
(a)
I
t has an eight-core, x86, 64-bit Jaguar CPU with a Radeon
GPU system-on-chip (SoC) from AMD with access to 8
Gbytes of RAM. It sports half a terabyte of hard-disk storage, a
Blu-ray drive, and a 3D image sensor. Of course, it has Gigabit
Ethernet, 802.11 b/g/n, USB 3.0, and HDMI output too.
So what is it?
It could be the Microsoft Xbox One or the Sony PlayStation
4 (see the figure). If you stand far enough back and don’t look
too closely at their specs, they look pretty much the same.
Differentiating the two platforms is more difficult, but the
developer’s job is significantly easier.
(b)
Microsoft’s Xbox One (a) and Sony’s PlayStation 4 (b) are very similar
architecturally, including the use of an eight-core SoC from AMD and
a Blu-ray disc drive.
There are some differences. Sony uses GDDR5, while
Microsoft went with DDR3. Sony’s optional 3D image sensor
is a pair of cameras, while Microsoft bundles the latest version
of the Kinect.
These platforms will be among the first to use AMD’s
heterogeneous Uniform Memory Access (hUMA) support
(see “Unified CPU/GPU Memory Architecture Raises The Per-
IMAGING TECHNOLOGIES
TAKE FLIGHT
NASA SUBJECTED ITS Interface
Imaging Spectrograph (IRIS)
observatory to vibration and
other testing before its June 26
launch. The satellite will use its
ultraviolet telescope and multichannel imaging spectrograph
to observe how solar material
moves, gathers energy, and
heats up as it travels through
the lower atmosphere. (courtesy
of LM Photo) 16
07.11.13
ELECTRONIC DESIGN
Gather, organize and control
data like never before.
Agilent 34970A/72A Data Acquisition Switch Units
(DAQs) have achieved market leadership for more
than a decade. Perhaps it’s the 34970A/72A DAQ’s
unmatched capability. Or the 34972A DAQ’s ability
to control everything remotely via mobile interface.
Or the BenchLink Data Logger software that lets
you scan, monitor, control and analyze your data so
you can quickly act on it. Maybe it’s all of the above.
349972A
972A DAQ
3-slot LXI unit with built-in 6½ digit DMM
USB and LAN
GPIB and RS-232
Grap
aphi
h cal web
interfacce
–
$1,9
$1
,933*
$1,6
,673
73**
Agilent and our
Distributor Network
Right Instrument.
Right Expertise.
Delivered Right Now.
© Agilent Technologies, Inc. 2013
* Prices are in USD and are subject to change
3497
34
970A
0A DAQ
Q
800-433-5700
www.alliedelec.com/agilent
Free $800* PC software upgrade
for advanced scanning control.
www.alliedelec.com/lp/agilentdaq/
News&Analysis
formance Bar” at electronicdesign.com),
which is part of AMD’s heterogeneous
system architecture (HSA). This puts
the GPU and CPU cores into the same
memory and cache space. Code that
runs on the GPU will still be distinct
from the CPU, but the tighter integra-
tion means less copying and quicker
dispatching. The similar SoC architectures will make a developer’s job much
easier.
The software that the two vendors
wrap around this hardware will be
much different, though. Microsoft is
One less hat to wear.
Let us be your power expert. We understand that you don’t
have the time to master every aspect of electronic design.
As a leading manufacturer of power supplies we are here to
collaborate with you to ensure your next project is a success.
Novum®
Advanced Power
Ac-Dc
Power Supplies
using Windows as the basis for the
Xbox One, but don’t expect it to run
Windows applications at this point. It’s a
gaming system, after all.
Power management is also critical to
these designs. The latest multiplayer,
3D, first-person shooter games will
tax the CPU and GPU cores. Yet these
cores mostly will be idle when users
are streaming videos or watching a
movie on Blu-ray.
Virtualization is likely to play
a central role as well. HSA makes
virtualization work across the CPU
and GPU cores. Games will be easier
to modularize, but security will be
improved. The vendors may open up
their systems a bit more by providing
a virtual machine for developers and
maybe even users so they have a bit
more flexibility in how they can use
the platforms.
PC gaming platforms will still be
able to deliver more computing horsepower, but the average PC or even a PC
designed for gaming is likely to have
less performance than these platforms.
Whether portability between platforms
will be improved remains to be seen,
but at least the hardware is there.
HSA likely will impact non-gaming
applications, but it now will have a
rather public pair of components. BILL WONG
Dc-Dc
Converters
www.cui.com/PowerExpert
18
07.11.13
ELECTRONIC DESIGN
Merger Propels Wired Home Networking
T
he recent merger of the HomePNA Alliance with the HomeGrid
Forum creates a new critical mass that
should be an incentive for increasing the
number of home area networks (HANs)
using wired technology. Today, most
networked homes use Wi-Fi for Internet access and video distribution.
HomePNA’s standard uses the frequencies above the standard DSL spectrum to carry video and Internet traffic.
AT&T’s U-verse IPTV services use this
technology. But with the availability
of the ITU-T’s G.hn wired technology,
equipment OEMs and service providers are slowly migrating to this flexible
standard, which supports phone-line
wiring, coax, or ac power-line media.
The HomeGrid Forum (HGF) was
created to support and champion G.hn,
while the HomePNA Alliance sought
to build and promote its own wired
home networking technology. Under
the HomeGrid Forum name, both will
perform compatibility and interoperability testing and certification as they
move to build the G.hn marketplace.
The HGF also will continue to support
and certify HomePNA products.
G.hn has been finalized for several
years now but has been slow to show
up in homes. The lack of certified sili-
con has been one problem, but so has
the general lack of consumer knowledge about the various types of home
network wiring potential. With most
homes having an installed base of ac
power line, telephone wiring, and cable
TV coax, no new wiring is needed. The
organizations will make a more consolidated effort toward G.hn’s promotion
and development. LOUIS E. FRENZEL
INTERFACES
AID AMPUTEES
THE DEFENSE ADVANCED Research
Projects Agency (DARPA) is
developing leaded implantable
myoelectric sensors to be used
as a novel peripheral-interface
technology with targeted muscle
re-innervation. These interfaces
use signals from nerves or muscles to control prosthetics and
provide direct sensory feedback.
(courtesy of DARPA) GO TO ELECTRONICDESIGN.COM
19
EngineeringFeature
JOE DESPOSITO | Editor in Chief
joe.desposito@penton.com
Survey Says
Design Engineers
Have Power On
Their Minds
A supercapacitor, two dc-dc converters,
and a charging coil garner four of the
top 10 spots on our annual list of
the Top 101 Components.
20
07.11.13 ELECTRONIC DESIGN
W
elcome to our yearly roundup of the Top
101 Components, compiling the most
sought devices in the electronics OEM
industry. We regularly receive mountains
of product announcements about connectors, power sources,
passives, LEDs, sensors, and more. Our editors then select the
most noteworthy for our design engineering audience.
We post these products on our Web site and promote them
via e-mail newsletters, tracking reader interest via Web site
analytics. This year, we also asked vendors to tell us what they
thought their best products were in the past 12 months. Readers then picked their favorite components of the ones we’ve
published and those that were nominated by the vendors.
Many design engineers are thinking about power and seeking more information about innovative power sources. In fact,
23 of the products on the list are categorized as power sources,
and four of the top 10 are power-related: the CLK series of
supercapacitors from Cellergy, the LXDC series of dc-dc converters for Murata Americas, the E200E series of dc-dc converters from MicroPower Direct, and the IWAS-3827EC-50
charging coil from Vishay Intertechnology.
SENSOR CLAIMS TOP SPOT
Though power sources made a strong showing, the component at the very top of the list is a sensor, the STMicroelectronics LIS3MDL. This is ST’s first discrete magnetometer, and it
sparked quite a bit of interest among our readers. Announced
in February, the LIS3MDL is now going to mass production.
One of the interesting applications of this standalone threeaxis magnetometer is indoor navigation, where it reliably
calculates dead-reckoning when no satellite signal is available.
The part also can be combined with other discrete sensors
from ST such as a three-axis micro-electromechanical systems (MEMS) accelerometer or three-axis MEMS gyroscope
to build sensors with as many as nine degrees of freedom.
ST has some indication of how this sensor might be used in
novel applications due to its experience with a 6x iNemo part,
an accelerometer plus magnetometer. Hapilabs, a company
based in Hong Kong, has developed a product called HAPIfork, which helps individuals monitor and track their eating
habits. With the help of indicator lights and gentle vibrations,
the device alerts users to the fact that they are eating too fast.
HAPIfork uses an ST accelerometer and a magnetometer
module, as well as an STM32 microcontroller.
LEDS TAKE SECOND PLACE
One of the hottest areas in electronics today is solid-state
lighting, so it makes complete sense that new LEDs would
f
c
d
e
b
a
Vishay Intertechnology’s IWAS-3827EC-50 is designed for the
wireless charging of 5-V portable electronics (a). Micro Power
Direct’s low-cost E200E dc-dc converters deliver 2 W of reliable
output power (b). Osram Opto Semiconductors’ latest automotive
LEDs offer high light output at high currents (c). Cellergy’s CLK
supercapacitors feature twice the capacitance of the company’s
general-purpose CLG series (d). Cree’s XLamp series LEDs provide discrete (XB-D) and multicolor (XM-L) LED options for applications ranging from architectural to vehicle and display lighting
(e). The standalone LIS3MDL three-axis magnetometer developed
by STMicroelectronics measures 2 by 2 by 1 mm (f).
GO TO ELECTRONICDESIGN.COM
21
EngineeringFeature
2013 TOP 101 COMPONENTS
Rank
Product
Company
Category
1
LIS3MDL magnetometer
STMicroelectronics
Sensors
2
XLamp series LEDs
Cree Inc.
LEDs
3
CLK series of supercapacitors
Cellergy
Power sources
4
LXDC2UR/LXDC3EP series
of dc-dc converters
Murata Americas
Power sources
5
E200E series of dc-dc converters
MicroPower Direct
Power sources
6
DCM4826X
stepper-motor drive module
Haydon Kirk Motion Systems
Motors & motion control
products
7
Oslon Black Flat LED
Osram Opto Semiconductors
LEDs
8
IWAS-3827EC-50 charging coil
Vishay Intertechnology
Power sources
22
9
T24-WSS wind-speed sensor
Mantracourt
Sensors
10
SBS75x family of PCB connectors
Anderson Power Products
Interconnects
11
Pi Sandwich case
Bud Industries
Cabinets & enclosures
12
A1250 bipolar switch
Allegro Microsystems
Switches
13
White Lite SSL connectors
Global Connector Technology
Interconnects
14
Displix LEDs
Osram Opto Semiconductors
LEDs
15
Nanominiature connectors
TE Connectivity
Interconnects
16
HyperLink PoE injector/midspans
L-com Inc.
Power Sources
17
2M series micro-miniature
connectors
Amphenol Aerospace
Interconnects
18
TSL4531 ambient
light sensor family
ams AG
Sensors
19
ESD-Safe MLC capacitor series
AVX Corp.
Circuit protection devices
20
ESD-SafeT X7R capacitors
AVX Corp.
Passive components
21
WM7121/WM7132 MEMS
microphones
Wolfson Microelectronics
Sensors
22
VCNL3020 proximity sensor
Vishay Intertechnology
Sensors
23
Transreflective TFT displays
Phoenix Display International
Displays
24
ULS series of dc-dc converters
Murata Power Solutions
Power sources
25
SPD500 pressure sensor
Sensirion AG
Sensors
26
AH1806/AH1808
Hall-effect switches
Diodes Inc.
Switches
27
RUW15 series of dc-dc converters
Murata Power Solutions
Power sources
28
MLO3 series capacitors
AVX Corp.
Passive components
29
F77 series of ultrasonic sensors
Pepperl+Fuchs
Sensors
30
TACmicrochip series capacitors
AVX Corp.
Passive components
31
MiniMax push-pull interconnect
Fischer Connectors
Interconnects
32
BFS 33M true color sensor
Balluff
Sensors
33
CSM series of
current-sense resistors
Stackpole Electronics Inc.
Passive components
34
ZWS10-30B series power supplies
TDK-Lambda Americas
Power sources
35
2Pro PPTC/MOV hybrid device
TE Circuit Protection
Circuit protection devices
36
BL3100C18650XSXPPGQA
intelligent battery pack
GlobTek Inc.
Power sources
pique our readers’ interest. In this particular case, the product that made the greatest
impression during the past year is the Cree
XLamp series of LEDs: the discrete (XBD) and multicolor (XM-L).
Built on Cree’s SC³ Technology Platform, the XB-D White LED delivers 139
lumens and 136 lumens per watt in cool
white (6000 K) or 107 lumens and 105
lumens per watt in warm white (3000 K),
both at 350 mA and 85°C.
Multicolor XM-L LEDs combine very
high efficacy at very high drive currents,
delivering 1000 lumens with 100 lumensper-watt efficacy at 3 A in a 5- by 5-mm
footprint. They’re designed for very-highlumen applications such as high-bay,
indoor commercial, or roadway lighting.
Cree wasn’t the only company that
placed an LED in our top 10. Osram Opto
Semiconductors secured the seventh
position with its Oslon Black Flat LED.
Designed for automotive front lighting
applications, the LED features high light
output at high currents, uniform light
distribution, thermal stability, and strong
contrast. It employs Osram’s UX:3 chip
technology, which is based on the company’s ThinGaN process. This technology uses a metallic mirror below its active
layer and a well-defined scattering surface
for optimized light extraction.
SUPERCAP LEADS POWER QUARTET
When Cellergy announced a supercapacitor with double the capacitance of its
CLG series last year, our readers noticed,
placing it third on our list. The CLK series
comes in two form factors, 17 by 17 and 28
by 17, and has an extended temperature
range from –40°C to 85°C. Capacitance
ranges from 30 to 240 mF.
Supercapacitors are generally used for
pulse applications and in conjunction with
batteries to assist in delivering energy to
the load. Cellergy plans on expanding the
CLK series to include additional voltages
and a version with lower leakage current.
Small dc-dc converters nailed the
next two spots. Murata Americas touted
its LXDC micro dc-dc converters as the
world’s smallest when they debuted in June
07.11.13 ELECTRONIC DESIGN
ANALOG
INTEGRATION
ISN’T FOR
EVERYONE
© 2013 Maxim Integrated Products, Inc. All rights reserved. Maxim Integrated and the Maxim
Integrated logo are trademarks of Maxim Integrated Products, Inc., in the United States and other
jurisdictions throughout the world.
EngineeringFeature
2013 TOP 101 COMPONENTS
Rank
Product
Company
Category
37
USB 2.0 connector
Conec
Interconnects
38
PXS series of power supplies
UltraVolt Inc.
Power sources
39
High-power TVS diodes
AVX Corp.
Circuit protection devices
40
CeeLok FAS-T connector
TE Connectivity
Interconnects
41
SMA6F series TVSs
Vishay Intertechnology
Power sources
24
42
PS25203 EPIC sensor
Saelig Co.
Sensors
43
PH9185NL transformers
Pulse Electronics
Power sources
44
Smart Position Sensor
Honeywell Sensing
and Control
Sensors
45
9286 series wire-to-wire
connector system
AVX Corp.
Interconnects
46
Illuminated tactile switch series
C&K Components
Switches
47
A1128 Hall-effect switch
Allegro Microsystems
Switches
48
RBE-12/20-D48 dc-dc converter
Murata Power Solutions
Power sources
49
PDJack connector
Molex
Interconnects
50
SM3KWxxA/SM5KWxxA series
TVS modules
ProTek Devices
Circuit protection devices
51
PwrBlade series connector
FCI
Interconnects
52
Okami OKX series of
dc-dc converters
Murata Power Solutions
Power sources
53
OA109AP/OA172SAP series
LED cooling fans
Orion Fans
(Knight Electronics)
Cooling products
54
Tekal series of
aluminum housings
Teko Enclosures
Cabinets & enclosures
55
JCK60 series dc-dc converter
XP Power
Power sources
56
DF50 connector
Hirose Electric Co. Ltd.
Interconnects
57
GCL-536 series switch
CW Industries
Switches
58
BD9521 battery disconnect switch
Gigavac
Switches
59
ERGO-CASE XS enclosure systems
OKW Enclosures Inc.
Cabinets & enclosures
60
NFC antenna
Pulse Electronics
Passive components
61
6U VPXtra 1000CD power supply
Behlman Electronics
Power sources
62
MK23 proximity switches
Meder Electronic Inc.
Switches
63
MPU-800S power supplies
MicroPower Direct
Power sources
64
15208A/25208A/35208A
digital accelerometers
API Technologies
Sensors
65
DDR4 DIMM sockets
FCI
Interconnects
66
Intelligent fan trays
Verotec
Cooling products
67
DH50 series
dc-dc power converters
Wall Industries
Power sources
68
AE458RFW series transformers
Coilcraft CPS
Power sources
69
GRN series power supplies
Power Sources Unlimited
Power sources
70
Shrouded male connectors
Mill-MaxX
Interconnects
71
Rig-Lok connector series
Amphenol Industrial
Global Operations
Interconnects
72
Bulk Metal Z1-Foil (FRST) resistors
Vishay Precision Group
Passive components
of 2012. The LXDC2UR has a 600-mA current rating and measures 2.3 by 2.5 mm,
while the LXDC3EP has a 1-A current rating and is 3.2 by 3.5 mm. Both combine
Murata’s proprietary materials and technical capabilities, specifically ferrite material expertise, multilayer processing, and
power module design. Murata has since
expanded the product lineup to accommodate higher voltage and current needs.
Fifth on the list, the MicroPower Direct
E200E series dc-dc converters come in a
four-pin system-in-package (SIP), a popular form factor with our readers. Six models operate from 5- and 12-V dc inputs and
have single outputs of 5, 12, or 15 V dc.
All deliver up to 2 W and are specifically
designed to provide reliable voltage conversion and/or power-line isolations for
space-critical board-level applications.
Rounding out the power sources quartet, Vishay Intertechnology’s IWAS3827EC-50 powdered-iron-based receiving coil targets the wireless charging of
5-V portable electronics. Compliant with
the Wireless Power Consortium, it measures 38 by 27 mm. It also is 44% smaller
and costs less than its predecessor, the
IWAS-4832FF-50, yet has equivalent performance and greater than 70% efficiency.
TINY STEPPER-MOTOR DRIVE, A
WIND-SPEED SENSOR, AND A PCB
CONNECTOR
S e c u r i ng s i x t h pl a c e i s t h e t i ny
DCM4826X stepper-motor drive module from Haydon Kerk Motion Systems.
The totally enclosed unit’s mounting plate
measures just 64 by 60 mm, with mounting-hole center-to-center distances of 52
by 38 mm. Total thickness is 34.4 mm.
Due to its small size, the DCM4826X is
well-suited for multi-axis stepper-motor
applications where space is at a premium.
As a bipolar, two-phase chopper drive, it
can be used with rotary stepper motors or
stepper-motor-based linear actuators. The
drive is integrated using the company’s
PDE User Interface via an RS-485 protocol
or a computer USB port.
Ninth on the Top 101 is a sensor that
measures wind speed and transmits the
07.11.13 ELECTRONIC DESIGN
EngineeringFeature
data wirelessly. Mantracourt’s T24-WSSPR-A can make high-accuracy measurements over a 5- to 125-mph range. Its
three-cup rotor is pressed on a stainlesssteel shaft with a Delron body fitted with
bronze Rulon bushings. The anemometer’s output value can be calibrated and
configured to the user’s requirements. On
the wireless side, the device powers down
between transmissions to maximize battery life in the field.
Like last year, just one interconnect
made it into the top 10. This time it’s
the SBS75x family of touch-safe, highpower connectors from Anderson Power
Products. The female part is a right-angle
connector that sits on the printed-circuit
board (PCB). This dual-pole connector
incorporates 105-A power contacts and
four auxiliary power contacts for applications requiring single or low-power capability up to 20 A.
The male signal contacts on the wire
side come in four lengths when sequenc-
2013 TOP 101 COMPONENTS
Rank
Product
Company
Category
73
BMA355 three-axis MEMS
accelerometer
Bosch Sensortec
Sensors
74
Sfernice P16S potentiometer
Vishay Intertechnology
Passive components
75
193 PUR-SI solar capacitors
Vishay Intertechnology
Passive components
76
P105 MiniTactor
Gigavac
Switches
77
OOSD101T1402-45TS TFT display
OSD Displays
Displays
78
ASR series resistors
Stackpole Electronics Inc.
Passive components
79
Mini reed relays
Meder Electronic Inc.
Relays
80
EMH-54/3-Q48N-C
dc-dc converter
Murata Power Solutions
Power sources
81
Non-contact proximity alarm
Standex Electronics
Transducers
82
Micro-SIM connector card
Molex
Interconnects
83
VLMU3100 UV LED
Vishay Intertechnology
LEDs
84
Miniature strain gauge
Vishay Precision Group
Sensors
85
FE series capacitors
AVX Corp.
Passive components
86
Family of ac-dc power supplies
Turck
Power sources
87
CCSO-914X SAW clock oscillator
Crystek Crystals
Timing devices
5FTU4PMVUJPOT
... from SRS !
2 GHz Signal Generator
SG382 ... $3900
10 MHz Rubidium Std.
FS725 ... $2695
Stanford Research Systems
Distribution Amplifiers
FS735 ... from $950
2 GHz Clock Generator
CG635 ... $2995
2013 TOP 101 COMPONENTS
Rank
Product
Company
Category
88
89
GLV-24-150 LED power supply
Green Watt Power
Power sources
SBT-MicroSD-01 socket
Ironwood Electronics
Interconnects
90
OV8835 CameraChip
OmniVision Technologies
Sensors
91
HPS series pressure switches
Honeywell Sensing
and Control
Switches
92
Microwave coaxial cables
Temp-Flex LLC,
a subsidiary of Molex Inc.
Interconnects
93
Series 770 NCC connectors
Binder-USA
Interconnects
94
COOL-CC3 chassis
Curtiss-Wright Controls
Defense Solutions
Cabinets & enclosures
95
FHD LCD module
NLT Technologies
Displays
96
Saf-D-Grid connector system
Anderson Power Products
Interconnects
97
9175/9176 series of connectors
AVX Corp.
Interconnects
98
BNO055 nine-axis sensor node
Bosch Sensortec
Sensors
99
MPA magnetic positioning sensor
Sick Inc.
Sensors
100
DF62 series of connectors
Hirose Electric Co. Ltd.
Interconnects
101
PPWA150B series
of power supplies
Power Partners
Power sources
FOR MORE INFORMATION
%XLOGLQJRQRXUULFK\HDUKLVWRU\RIGHVLJQLQJKLJK
SHUIRUPDQFHVFLHQWL¿FDQGHQJLQHHULQJWHVWHTXLSPHQW
656QRZRIIHUVDEURDGOLQHRI5)LQVWUXPHQWDWLRQ
656SURGXFWVDUHQRWRQO\DIIRUGDEOHEXWWKH\RIIHU
LQFUHGLEOHSHUIRUPDQFH²OLNHWKHQHZ6*
'&WR*+]6LJQDO*HQHUDWRUZKLFKKDVRXWVWDQGLQJ
Frequency Counter
SR620 ... $4950
ing is required. They also provide a last
mate/first break connection in relation to
the power contacts. The touch-safe mating interface protects the user by eliminating finger contact with live circuits
(per UL1977, section 10.2). The device
saves space and increases design flexibility by eliminating the need to bring wires
to the PCB. Press-fit containers on the
housing secure and position the connector to the PCB prior to soldering.
This wraps up our look at the top 10
components on this year’s list. We’d also
like to congratulate all of the companies
whose innovative products made this
year’s Top 101. You can find the full roster
of components in the table. The online
version of the table contains links to the
full product reports on electronicdesign.
com, where you can find links to the manufacturers’ Web sites and to many of the
datasheets for these products.
VSHFL¿FDWLRQVDQGVHOOVIRUDERXWWKHSULFHRI
FRPSHWLQJLQVWUXPHQWV
3OHDVHYLVLWXVDWZZZWKLQN656FRPRUFDOOXVDW
IRUGHWDLOVRQWKHVHSURGXFWVDVZHOO
DVRXUIXOOOLQHRILQVWUXPHQWDWLRQ
30 MHz Function/Arb Generator
DS345 ... $1595
4 GHz Signal Generator
SG384 ... $4600
10 MHz Rb Oscillator
PRS10 ... $1495
6 GHz Signal Generator
SG386 ... $5900
5FM
tXXXUIJOL434DPN
TechnologyReport
LOUIS E. FRENZEL | Communications Editor
lou.frenzel@penton.com
Wi-Fi
And Bluetooth Rule
The Airwaves
These two core wireless technologies continue to dominate short-range applications
with new features and benefits.
W
ith dozens of short-range wireless standards to choose from,
engineers still defer to Bluetooth
and Wi-Fi. After more than 15
years, they continue to grow, improve, and provide new
features and benefits. These amazingly useful technologies both use the 2.4-GHz industrial-scientific-medical
(ISM) unlicensed spectrum. Wi-Fi also has access to the
5-GHz and 60-GHz ISM spectrum.
WHAT’S NEW WITH BLUETOOTH?
The most recent version of Bluetooth, 4.0, incorporates all
of its previous features but adds Bluetooth Low Energy (BLE).
Also known as Bluetooth Smart, BLE uses a different set of
technical and radio techniques to ensure very low power consumption compared to the original Bluetooth standard (see
“Bluetooth’s Background” at electronicdesign.com).
BLE’s data protocol was changed to support low-duty-cycle
transmissions or very short transmission bursts between long
periods. The low duty cycle and extremely low-power sleep
modes enable BLE products to operate for many years on a coin
cell. BLE also provides a group of application programming
interfaces (APIs) for fast and easy application development.
28
BLE still operates in the same ISM license-free 2.4- to 2.483GHz frequency band as standard Bluetooth, but it uses a different frequency-hopping spread-spectrum (FHSS) scheme.
Standard Bluetooth hops at a rate of 1600 hops per second over
79 channels that are 1 MHz wide. BLE FHSS uses 40 channels
that are 2 MHz wide to ensure greater reliability over longer
distances. Standard Bluetooth offers gross data rates of 1, 2, or 3
Mbits/s. BLE’s maximum rate is 1 Mbit/s with a net throughput
of 260 kbits/s.
Also, BLE uses Gaussian frequency shift keying (GFSK)
modulation. It offers a power output of 0 dBm (1 mW) and a
typical maximum range of 50 meters. Security is via the 128-bit
Advanced Encryption Standard (AES). An adaptive frequency-hopping technique that avoids interference, a 24-bit cyclic
redundancy code (CRC), and a 32-bit message integrity check all
improve link reliability. The most common network configurations are point-to-point (P2P) and star. Latency is only 6 ms.
Most low-power wireless networks are involved in consumer, medical, health, or sports and fitness applications. These
mobile applications require small size and coin cell power that
can function for years without attention.
Consumer applications include electronic leashes to locate
people, animals, or things; proximity detection for use in iden-
07.11.13
ELECTRONIC DESIGN
GO TO ELECTRONICDESIGN.COM
29
TechnologyReport
tification, authentication, and wireless locks; RFID-like
(radio-frequency identification) cases; automatic meter
reading; toys; automotive applications such as On-Board
Diagnostic II connector plug-ins; home-area networks;
human interface device (HID) peripherals like mice and
keyboards; and game controllers.
Medical and health care applications include heart-rate
monitors, temperature monitors, instrumentation, and
body-area networks. Similarly, sports and fitness applications include sports watches and monitors, heart-rate
belts, bike computers, speed and distance monitoring, and
fitness equipment.
BLE isn’t compatible with standard Bluetooth, though.
It is a separate radio on standard Bluetooth 4.0 chips. BLE
devices do not interoperate with classical Bluetooth. But
such interoperability could be implemented with a dualmode IC that integrates a standard Bluetooth radio and a
BLE radio, operating separately but not at the same time.
They also can share an antenna. Dual-mode devices are
available separately for low-power-only applications.
Bluetooth Smart Ready refers to the dual-mode chips
of version 4.0. It handles previous versions of classic Bluetooth in addition to BLE. Bluetooth Smart is BLE only. It
won’t connect with any other version of Bluetooth except
for Smart Ready 4.0 or other BLE devices. Classic Bluetooth is compatible with previous versions and Smart
Ready devices. It won’t connect with Smart devices.
Recent announcements from Apple, Google Android,
and Microsoft (Windows 8) indicate software support for
the new Bluetooth standards. This support will expedite
connectivity with Smart devices to smart phones, tablets,
and laptops.
Multiple vendors offer dual-mode and BLE chips,
including Broadcom, CSR, Dialog Semiconductor, EM
Microelectronics, Nordic Semiconductor, Texas Instruments, and the TTP Group. Nordic Semiconductor, for
example, offers the nRF51822 BLE radio (Fig. 1).
WI-FI REVIEW
Wi-Fi is the trade name of the IEEE’s wireless local-area
network (WLAN) standard, designated 802.11. A suffix
at the end indicates the standard version. 802.11n is the
current most widely implemented variant found in Wi-Fi
access points, hotspots, and routers. Wi-Fi is also in most
of today’s laptops, all tablets and smart phones, and smart
television sets.
The 802.11 standard defines both a physical layer
(PHY) and a media access control (MAC) layer in the networking scheme. While the MAC tends to remain mostly
the same, the PHY changes to include the most recent
wireless technology for greater speed and link reliability.
30
1. Nordic Semiconductor’s nRF51822 Bluetooth Low Energy
module can be embedded in almost any application. The radio
features data rates of 250 kbits/s, 1 Mbit/s, and 2 Mbits/s with
an output power of +4 dBm and a receive sensitivity of –92.5
dBm. The module incorporates an ARM Cortex M0 CPU with 256
kbytes of flash and 16 kbytes of RAM.
The first generation 11b showed up in 1997 and uses
direct sequence spread spectrum (DSSS) to achieve data
rates to 11 Mbits/s in a 20-MHz channel in the 2.4-GHz
ISM radio band. With the growth of the Internet, that low
speed soon became a disadvantage.
The second-generation 802.11a appeared in 1999. It was
the first to use the 5-GHz ISM band and orthogonal frequency division multiplexing (OFDM) with 64 subcarriers
spaced 312.5 kHz apart. Channel bandwidth was 20 MHz,
and binary phase-shift keying (BPSK), quadrature phaseshift keying (QPSK), 16-phase quadrature amplitude
modulation (16QAM), and 64-state quadrature amplitude
modulation (64QAM) types were defined, permitting the
data rate to increase to a maximum of 54 Mbits/s.
While the 802.11a version was more robust because of
the OFDM characteristics that mitigated multipath reflections and the 5-GHz assignment meant less interference,
the higher frequency still limited the range. This version
was never too popular despite its advantages.
The big breakthrough came when the 802.11g standard
was approved in 2003. It is essentially the same as 11a but
operates in the 2.4-GHz band. Using the same OFDM and
modulation options, it too can deliver up to 54 Mbits/s. It
was immediately popular because the many IC companies
competing for the business brought chip prices very low.
The current 11n standard is a further improvement
over 11a/g. It adds 40-MHz channels and multiple-input
multiple-output (MIMO) features to the OFDM, allowing
data rates to increase to as much as 600 Mbits/s.
MIMO uses multiple receivers, transmitters, and
antennas to achieve spatial division multiplexing (SDM).
SDM transmits fast multiple data streams concurrently
07.11.13
ELECTRONIC DESIGN
Generate the exact waveforms you want.
Introducing Agiileent Trueform
m technology.
Agilent 33500B Series waveform generators
with exclusive Trueform signal generation
technology offer more capability, fidelity
and choice than traditional DDS generators.
Get the precision and flexibility you need to
characterize new components and designs
with superior confidence. And accelerate
development from start to finish.
Agilent and our
Distributor Network
Right Instrument.
Right Expertise.
Delivered Right Now.
© Agilent Technologies, Inc. 2013
800-433-5700
www.alliedelec.com/agilent
33500B Series Waveform Generators
Sine waves with 5x lower harmonic distortion
Pulses up to 30 MHz with 10x less jitter
Point-by-point arbitrary waveforms with sequencing
16 -bit resolution with 1 mVpp to 10 Vpp amplitude
8 upgradable models in 20 & 30 MHz
Learn more about the advantages
of Trueform waveform generation
www.alliedelec.com/lp/agilentfunctiongenerators
TechnologyReport
within the same 20- or 40-MHz channel bandwidth. Precoding and post-decoding as well as unique path characteristics distinguish the data streams. The data rate then
can be multiplied by a factor roughly equal to the number
of data streams.
The 11n MIMO standard permits up to four transmit and four receive channels (4x4), although 1x2, 2x2,
and 3x3 versions are more widely used. The 600-Mbit/s
data rate is achieved using 4x4 MIMO with 64QAM in a
40-MHz channel.
What makes 11n so popular is its ability to carry video,
enabling wireless connectivity in TV sets, DVD players, and other video equipment. Using compressed video
(H.264, MPEG4, etc.), full HD signals can be transmitted
reliably with the higher levels of MIMO.
The WLAN space is dominated by 11n Wi-Fi, which is
available in all smart phones, tablets, and laptops. It also
is the wireless technology of most hotspots and access
points, including millions of home wireless routers. It’s
increasingly being embedded in other consumer electronics. And, it’s backwards compatible with previous standards, allowing 802.11a/g equipment to be used.
32
2. The Texas Instruments SimpleLink Wi-Fi CC3000 BoosterPack development board and MSP430 LaunchPad evaluation kit
speed the creation of Wi-Fi based IoT applications.
The latest version of the standard is 802.11ac. It isn’t
fully ratified yet, but it’s essentially complete. Finalization
is expected in 2014. It uses the 5-GHz ISM band only for
minimum interference and maximum available bandwidth. Furthermore, it continues the use of MIMO and
OFDM. Some key changes boost theoretical
data rate in excess of 3 Gbits/s, depending
on modulation, channel bandwidth, and
MIMO configuration. The primary changes
are its 80- and 160-MHz wide channels in
addition to the usual 40-MHz channel.
As the bandwidth increases, so does the
number of OFDM subcarriers to a maximum of 512 at 160-MHz bandwidth. OFDM
also adds 256QAM, which further boosts
data rate. And, 11ac defines a greater number of MIMO versions with a maximum of
an 8x8 configuration. A multi-user version,
MU-MIMO, is defined as well. The standard additionally supports coexistence and
compatibility with previous 11a and 11n
devices. Transmit beamforming is an option
to extend range and ensure link reliability.
Products for 11ac are emerging. Routers and access points are showing up, but
few user devices include it. But because of
its many advantages in speed and freedom
from interference, you will eventually see
11ac capability not only in laptops but also
tablets and smart phones.
Another recent version of Wi-Fi,
802.11ad, uses the 60-GHz ISM band. It
is backwards compatible with all previous
versions, including 11a/b/g/n/ac, as the
07.11.13
ELECTRONIC DESIGN
Technical Notes
The Avago Advantage
Driving and Protecting IGBTs in Inverter
BIPOLAR
BUFFER
Gate Drive
Optocoupler
IOUT
The ACPL-339J is an advanced 1.0 A dual-output, easy-to-use, intelligent
IGBT gate drive optocoupler interface. Uniquely designed to support
MOSFET buffer of various current ratings, the ACPL-339J makes it easier
for system engineers to support different system power ratings using one
hardware platform by interchanging the MOSFET buffers and power IGBT/
MOSFET switches. This concept maximizes gate drive design scalability for
motor control and power conversion applications ranging from low to high
power ratings.
The ACPL-339J is also integrated with short circuit protection, under
voltage lockout (UVLO), “soft” IGBT turn-off, and isolated fault feedback to
provide maximum design flexibility and circuit protection.
13
VCC2
12
VOUTP
UVLO_P
CATHODE
%
3
*
7
&
3
3
2,4
LED1
SHIELD
Drive Logic
& Overlap
Protection
16
9
6
LED2
15
7
DESAT
VGND1
VE
UVLO_N
11
VCC1
FAULT
IGBT
1200V/50A
IOUT
0.4A
Introduction
ANODE
3A
14
VOUTN
VEE
Gate Drive
Optocoupler
IOUT
8A
IGBT
1200V/300A
IOUT
1A
Gate Drive
Optocoupler
IOUT
16A
IGBT
1200V/600A
IOUT
3A
Figure 2: Gate drivers with different output currents to match
BIPOLAR buffers
of the BIPOLAR’s output current is a factor of its base current, IB and the
transistor current gain, β. In another words, gate drivers with different
peak output currents and matching BIPOLAR buffers are needed to achieve
the peak gate current required by different class of IGBTs.
MOSFET buffers on the other hand are voltage controlled devices and their
current amplifications are independent on the previous gate driver stage.
Figure 3 shows MOSFET buffers with different internal turn-on resistance,
RDSON to deliver the peak gate current required by the different class of
IGBTs. Although the ACPL-339J output is specified at 1A, the switching of
the MOSFET buffer will happen as long as the ACPL-339J output voltage
crosses the input threshold of the MOSFET buffer.
DESAT
VGMO3
MOSFET
BUFFER
5,8
SHIELD
3A
Figure 1: Functional diagram of the ACPL-339J gate drive optocoupler
IGBT
1200V/50A
ACPL-339J
Driving IGBTs
8A
IGBT
1200V/300A
One of the key feature of gate drive optocoupler is its ability to provide high
peak output current to charge or discharge the gate of the IGBT quickly to
prevent switching loss. Avago’s gate drive optocoupler has output current
ranging from 0.4A up to 5A which can be used to drive small IGBT directly.
For IGBT with higher ratings, discrete PNP/NPN bipolar buffer stage is
usually used. By changing the buffer stage to MOSFET, it maximizes gate
drive design scalability and power conversion efficiency.
Figure 2 shows different BIPOLAR buffers are used to drive different class of
IGBTs from 50A to 600A. As the size of IGBT gets bigger, higher peak current
is required at the gate of the IGBT to turn it on efficiently. The magnitude
VOUT
Signal
16A
IGBT
1200V/600A
Figure 3: ACPL-339J to drive MOSFET buffer with different output
current capability
Your Imagination, Our Innovation
Sense • Illuminate • Connect
The Avago Advantage Technical Notes
The ACPL-339J makes it easier to support different system power ratings
using one gate drive platform by interchanging the MOSFET buffers and
power IGBT/MOSFET switches. And all these changes can be made without
redesigning the critical circuit isolation and short circuit protection.
Protecting IGBTs
BIPOLAR
BUFFER
Gate Drive
Optocoupler
The IGBT’s collector-emitter voltage (VCE) can be monitored by ACPL339J DESAT pin during IGBT normal operation. When short circuit
occurs, high current flows through the IGBT and it comes out of
saturation into DESAT mode. This causes the IGBT’s VCE to increase
rapidly from a saturation voltage of 2V. Once its crosses the ACPL-339J’s
threshold of 8V, a short circuit fault is registered and soft shutdown
is triggered. The ACPL-339J’s VGMOS pin will switch on an external
transistor to slowly discharge the gate of the IGBT to achieve the soft
shutdown effect. The rate of the soft shutdown can be adjusted by the
size of the external transistor and resistor to minimize the overshoot
at the IGBT. And lastly, the entire DESAT operation is completed by
reporting the FAULT through a built-in insulated feedback path to
the controller.
PPEAK=32W
VCESAT=2V
IGBT
1200V/600A
16A
IOUT
3A
MOSFET
BUFFER
Gate Drive
Optocoupler
PPEAK=2W
RDSON=8mΩ
VOUT
Signal
IGBT
1200V/600A
16A
Figure 4: MOSFET buffer consumes less power and improves the
overall efficiency
Summary
BIPOLAR buffer uses a compounded structure consisting of 2 or more
transistor stages cascaded to achieve high current gain. The drawback
is the increased in saturation voltage, VCESAT and the output not able
to pull to the rail. This results in high power losses when the BIPOLAR
buffer delivers high peak output current to the gate of the IGBT. The
power loss from the BIPOLAR buffer increase tremendously as higher
peak current is needed by the gate of the IGBT.
VC 16
1 NC
R
+
–
+
–
2 CATHODE
3 ANODE
R
0.3μF RF=10k Ω
CF=1nF
4 CATHODE
MOSFET buffer has “rail-to-rail” output and lower internal turn-on
resistance, RDSON while delivering higher peak current compared to bipolar
buffer. The MOSFET buffer in figure 4 shows significant power reduction
when compared to the BIPOLAR buffer delivering the same peak current.
DESAT 15
VOUTP 12
6 VCC1
VOUTN 11
7 FAULT
NC 10
8 VGND1
VCC
4
100 Ω
DDESAT
330 Ω
1μF
2 VGMOS 14
VCC2 13
5 VGND1
CBLANK
1μF
The concept of ACPL-339J driving MOSFET buffer helps to maximize
gate drive design scalability from low to high power systems. This helps
to reduce design cycle time and together with the integrated short
circuit protection and feedback, the gate drive design is simplified with
less PCB space and costs.
^
RP
R3=330 Ω
+
MP1 –
RGP
MN1
RGN
+
MN2
^
RN
+
10μF
+
–
9
+
VCE
–
Q1
1
Short
Circuit
MN3
3
MP1=Si7415DN / Si7465DP
MN1=Si7414DN / Si7848DP
Q2
8V
1
• Short circuit causes IGBT to go into DESAT and VCE increases
• DESAT pin detects FAULT and triggers soft shutdown
2
• VGMOS switches on MN2 for soft shutdown to prevent VCE
voltage overshoot due to parasitic inductance
VG(IGBT)
3
• The size of MN2 and RS can be adjusted to achieve desired
soft shutdown slew rate
FAULT
4
• FAULT pin reports DESAT fault
VDESAT
DC BUS
VGMOS
Soft Shut Down
L
IGBT
V=L di/dt
hard shutdown during short circuit will cause
high VCE overshoot stressing the IGBT
Figure 5: Short circuit protection, “soft” IGBT turn-off, and isolated fault feedback
Contact us for your design needs at:
Short
Circuit
www.avagotech.com/gatedriver
Avago, Avago Technologies, the A logo are trademarks of Avago Technologies in the United States and other countries.
All other trademarks are the property of their respective companies.
Data subject to change. Copyright © 2013 Avago Technologies
MAC layers of the protocol are similar.
The 11ad version is also known by its
trade name, WiGig. The Wireless Gigabit
(WiGig) Alliance, the trade association
supporting and promoting 11ad, recently
merged with the Wi-Fi Alliance (see “The
Wi-Fi Alliance Ensures Compatibility” at
electronicdesign.com).
WiGig us es the unlicens ed ISM
60-GHz band from 57 to 64 GHz, dividing it into four 2.16-GHz bands. Its
OFDM primary modulation scheme can
support up to 7 Gbits/s, making it one of
the fastest wireless technologies available.
A single-carrier mode uses less power
and is a better fit for some portable handheld devices, delivering up to 4.6 Gbits/s.
Uncompressed video can be transmitted
at both speeds. The WiGig specification
also provides AES security.
Because of the small antenna size at 60
GHz, gain antennas are normally used to
boost signal power and range. The maximum typical range is 10 meters. WiGig
products use antenna arrays that can provide beamforming. This adaptive beamforming permits beam tracking between
the transmitter and receiver to ensure that
obstacles can be avoided and speed maximized even under changing environmental conditions.
One clever feature of the standard is its
use of a protocol adaption layer (PAL).
This software structure talks to the MAC
layer and allows simplified wireless
implementation of other fast standard
interfaces like USB, HDMI, DisplayPort,
and PCI Express.
Wi-Fi continues to be one of the most
widely used wireless technologies in history. More than 1 billion Wi-Fi enabled
devices were shipped in 2011, and those
shipments are expected to double by 2015.
Many new devices in the works are related
to consumer media products, health/fitness/medical, automotive, smart meters,
and automation products.
WHAT’S NEW WITH WI-FI?
The IEEE 802.11 Working Groups keep
a running agenda of development. They
work on the next major standard versions
as well as many additions and improvements to the basic standards. The most
significant standards include:
• 802.11e: provides quality-of-service
(QoS) features that allow Voice over
Internet Protocol (VoIP) and other critical services to be carried over Wi-Fi
• 802.11i: provides full security for Wi-Fi
in the form of Wired Equivalent Privacy (WEP), Wireless Protected Access
(WPA), and Wi-Fi Protected Access 2
(WPA2)
• 802.11s: brings automatic ad hoc mesh
networking to Wi-Fi
• 802.11u: a protocol between access
points and clients that permits internetworking with support for authentication, authorization, and accounting
with network selection, encryption policy enforcement, and resource management; facilitates automatic connections
and network handoffs
• 802.11y: brings Wi-Fi to the 3650- to
3700-MHz band
One of the most interesting new versions is the 802.11p standard, which will
be deployed in V2x or vehicle-to-vehicle (V2V) and vehicle-to-infrastructure
(V2I) systems. These systems are part of
the proposed Intelligent Telecommunications System (ITS), which provides safety
and traffic information to vehicles using
telematics. Called the Dedicated Short
Range Communications (DSRC) wireless system, this standard is similar to the
802.11a Wi-Fi standard.
The Dedicated Short-Range Communications (DSRC) system uses the IEEE
802.11p standard and a protocol known
as Wireless Access in Vehicular Environments (WAVE). The DSRC is assigned a
75-MHz segment of spectrum from 5.85
to 5.925 GHz. There are seven 10-MHz
channels designated by even numbers
from 172 to 184. These channels are half
the size of a standard 20-MHz 802.11
channel to minimize Doppler shift and
multipath fading. A European version of
42+/*
3
.
,
5
+
0+4/./"
./"4&
5)35114
!5./1/,2
25
"+-*
0
)3"(.*3
*.024()
.
4
00
+
)
214,
1 - 2
334&(- * 5 2 / (-
# # #
(- 2 3 03
2#4&52,(034,52)
+"(24!5./1,25)351
Impedance Levels 10 ohms to 250k ohms,
Power Levels to 3 Watts, Frequency Response
±3db 20Hz to 250Hz. All units manufactured
and tested to MIL-PRF-27. QPL Units available.
&2#354
44/"+-*251
Ideal for Noise, Spike and Power Filtering
Applications in Power Supplies, DC-DC
Converters and Switching Regulators
&+0134!5./1,25)351
10 Nanoseconds to 100 Microseconds.
ET Rating to 150 Volt Microsecond,
Manufactured and tested to MIL-PRF-21038.
+0*(034.*.4+1
&+0134!5./1,25)351
Plug-In units meet the requirements
of QPL-MIL-PRF 21038/27.
Surface units are electrical equivalents
of QPL-MIL-PRF 21038/27.
% 4 2/35*35
!5./1,25)351
Input voltages of 5V, 12V, 24V And 48V.
Standard Output Voltages to 300V (Special
voltages can be supplied). Can be used as self
saturating or linear switching applications. All
units manufactured and tested to MIL-PRF-27.
$''''4
&2#354!5./1,25)351
0.4 Watts to 150 Watts. Secondary Voltages 5V
to 300V. Units manufactured to MIL-PRF-27
Grade 5, Class S (Class V, 1550C available).
ek
ne we
k to o ies
c
o
t
S
it
ry
nt
Delive ample qua
for s
.004*2004,5334''%$%'$
$%%$''
$%%
& 03-*52/(-14/-
143 Sparks Ave. Pelham, N.Y. 10803
www.picoelectronics.com
GO TO ELECTRONICDESIGN.COM
5)4XL]
TechnologyReport
,IRQO\5)FRXOG
EHVRHDV\
5)0RGXOHV
5HPRWH&RQWUROV
$QWHQQDV
5)&RQQHFWRUV
&XVWRP'HVLJQV
the system is assigned 50
MHz of bandwidth for
five channels.
The spectrum permits vehicle and roadside unit (RSU) infrastructure radios to form
vehicular ad hoc networks (VANETs). These
dynamic networks will
p r ov i d e s h o r t - l i v e d
intermittent connectivity to implement various safety applications 3. The Ruckus Wireless 7782 outdoor carrier-grade access point
for collision avoidance uses 30° and 120° antenna patterns to improve link reliability,
and road safety. Connec- capacity, and speed.
tions are automatic. The
802.11p standard radio is half duplex with are least likely to interfere with the TV
a data rate in the 6- to 27-Mbit/s range channels. A national database of TV staand has an estimated maximum range of tion locations and open channels lets
the cognitive radio hop to an appropri300 meters (1 km).
The WAVE protocol uses the standard ate channel. You will hear of 802.11af
PHY and MAC layer of 802.11a but uses referred to as Super Wi-Fi or White-Fi.
The 802.11af standard defines the PHY
IEEE 1609 for the upper layers including
a MAC extension, a logical link control and MAC layer for white space operation.
(LLC) layer, network and transport layers While the standard is still under developthat include IPv6 with User Datagram ment, the basic features are defined. It
Protocol (UDP) and Transmission Con- uses the TV channels in the 470- to 710trol Protocol (TCP), and an upper mes- MHz range, or UHF TV channels 20 to
sage application layer. The 1609.2 stan- 31. The channel bandwidth is 6 MHz. The
dard also provides security. Several types basic modulation scheme is OFDM with
of short message formats have been devel- BPSK, QPSK, 16QAM, or 64QAM. The
transmit power level is 20 dBm, and the
oped for different conditions.
Then there are projects still being antenna gain is 0 dBi.
These low frequencies offer exceptiondeveloped by the Task Groups, such as
802.11ah. This modification to exist- al range, far greater than other Wi-Fi vering standards would allow operation in sions in the 2.4- and 5-GHz range. Ranges
sub-1-GHz bands to extend the range to several miles are possible. Data rates
and implement applications such as will vary depending on modulation type,
smart meters. Another new standard in range, and other factors. Equipment for
development, 11af, targets the TV white white spaces initially will use proprietary
space opportunity. It seeks to meet the wireless standards. The 802.11af standard
legal requirements for channel access and will be used with these early systems as
well as another IEEE standard, 802.22.
coexistence in the TV white space bands.
White space refers to the unused TV
channels that are present around the U.S. A WI-FI FUTURE
The idea is to use the vacant channels for
While Wi-Fi appears to be everywhere,
short-range data transmission. The num- soon it will be even more widespread. It
ber and location of these unused channels is already available on many airlines so
vary widely, so cognitive radio techniques passengers can connect to the Internet
are used to identify vacant channels that in flight. It is used in printers and cam-
ZZZOLQ[WHFKQRORJLHVFRP
07.11.13
ELECTRONIC DESIGN
eras. But that’s not all. Wi-Fi also is targeting the machine-tomachine (M2M) field and the Internet of Things (IoT).
Wi-Fi appears to be emerging as the wireless technology
of choice in home networks and even appliances. ZigBee and
Z-Wave wireless devices are already in the home, and they
require some kind of gateway to connect to the Internet. Since
many homes have a Wi-Fi network, it is a natural choice for
home networks. Whirlpool, LG, and other companies use Wi-Fi to collect usage
data on their refrigerators, washers, and
other appliances.
Texas Instruments designed its SimpleLink Wi-Fi CC3000 BoosterPack to
help engineers develop Wi-Fi enabled
IoT products. It offers simplified Wi-Fi
connectivity for microcontroller-based
systems and works with TI’s MSP430 and
TivaC series for MCUs. The TI CC3000
module is a self-contained wireless network processor that simplifies the implementation of Internet connectivity (Fig. 2).
Wi-Fi als o is incre asing ly b eing
deployed in factory automation, petrochemical plants, and other challenging RF
environments. Enterprise-class throughput is combined with a broad operating
temperature and rugged packaging to
provide a reliable wireless access point for
industrial users.
Carrier-grade Wi-Fi refers to ruggedized access points (APs) with special
performance characteristics. The Ruckus
Wireless ZoneFlex 7782 outdoor AP features 802.11n operation in the 2.4- and
5-GHz bands (Fig. 3). Its internal smart
antennas provide 120° and 30° patterns for
improved capacity and performance.
The latest and most ambitious use of
Wi-Fi is in offloading data from cellular
networks. Cellular carriers are trying hard
to handle the big explosion of data, like
video to smart phones and tablets. The
ongoing upgrade to 4G LTE is going well
but takes time and capital investment.
In many instances, the networks are not
keeping up with the demand for speed.
Wi-Fi networks are being implemented to handle the load. When a cellular
connection cannot handle the speed, the
user is passed off to a nearby Wi-Fi network if one is available. Many cellular
GO TO ELECTRONICDESIGN.COM
carriers are adding or expanding their own Wi-Fi networks to
ensure that the offload process can occur seamlessly. The
offload is not fully automatic, but work is in progress to make
that happen. Carriers are discovering that Wi-Fi offload provides cheaper and faster service than LTE. Look for more
hardened, carrier-grade Wi-Fi as cellular providers implement the offload option.
35
EngineeringEssentials
BILL WONG | Embedded/Systems/Software Editor
bill.wong@penton.com
Advanced Packaging
Delivers Capacity
AND PERFORMANCE
High-density BGAs and 3D stacking increase
compute and storage density.
INSIDE THE CHIP
The die inside a chip are much
smaller than the chip. A lead
frame typically surrounds the
die. Early microcontrollers
were delivered in packages
as large as 40 pins like the
8-bit Intel 8742 (Fig. 1). The
2. Intel’s many Pen- die has a large bonding pad.
t i u m g e n e r a t i o n s A bonding wire connects
utilized a multilayer the pad to a lead in the lead
package to link the frame that is the pin on the
bonding wires to the exterior of the chip.
DIPs limited the size of
I/O pins.
pins, but the approach still
works for some surface-mount packages. For example, flat-pack packages are internally similar to the
technology used on the 8742, but the exterior pins
1. The Intel 8742 was an 8-bit microwere designed for surface mounting. Dual and quad
controller that came in a 40-pin DIP.
flat-pack (QFP) chips are now very common for
(photograph by Ioan Sameli)
surface-mount applications.
Leadless chip carriers (LCCs) essentially eliminate
the
leads
but keep the same internal architecture. The
hip die sizes continue to grow, as does the number of
connections
are
exposed on the bottom and sometimes the
connections. The switch to high-speed serial interside
of
the
chip.
These
chips allow denser circuit boards. Some
faces reduced the required pin count, but that simply
LCC
form
factors
can
be
mounted in matching sockets. These
slowed the pin count down.
form
factors
typically
were
processors or more expensive
Dual inline packages (DIPs) used to be the mainstay for the
components.
electronics industry, and many applications still use them. They
A single lead frame does not work well with chips that
worked well with sockets and circuit boards. Through-hole
sprout
hundreds of pins. A move to stacked frames like the
connections allowed double-sided and then multilayer boards.
ones
used
in Intel’s Pentium allows more bonding wires to be
The problem is that DIPs hit limitations as they scaled
connected
in a smaller area (Fig. 2).
down in size and the number of connections rose. These days,
Ball
grid
arrays (BGAs) place the chip on top of a doublesurface-mount technology (SMT) has become the primary
sided
circuit
board (Fig. 3). The bonding wires on top of the
way to utilize electronic components.
C
36
07.11.13 ELECTRONIC DESIGN
Bonding wire
Die
Metal layer
Circuit board
Solder balls
3. A BGA has solder balls on the bottom of the circuit board. The
die sits on top, and the bonding wires connect the die to the board.
Through holes connect the top layer to the bottom.
Redistribution layer metal
Dielectric layer 2
Solder ball
Dielectric layer 1
Die
Aluminum bond pad
Die second passivation layer
4. Freescale’s Kinetis KL02 wafer-scale package has the solder balls
attached directly to the die.
die connect to a small, double-sided circuit board. Solder balls
are attached to the bottom of the board, and through-holes
link the balls to the pads connected to the gold bonding wires.
CHIP AND WAFER SCALE PACKAGING
BGAs are moving toward chip-scale packaging (CSP),
where the die size is very close to the package size. In some
cases there are cooling and power challenges.
Wafer-scale packaging eliminates the BGA circuit board and
attaches the solder balls directly to the die (Fig. 4). Freescale’s
Kinetis KL02 uses this approach. This 48-MHz, 32-bit Arm
Cortex-M0+ microcontroller fits into a 1.5-mm by 2-mm
package (Fig. 5). The chip is so small, it can be used in medical
devices that would be swallowed. Of course, using only 36 μA/
MHz helps keep power requirements down as well.
where multiple die can be blended in one package. We will
take a look at 3D die later. 3D tends to be a much different
approach than simply packaging die together.
Multichip carriers enable designers to put more than one
die into the package. There are numerous multichip solutions,
but in general they employ a circuit board to connect the chips
together and to link them to the pins or solder balls.
Multichip carriers give designers a common pin-out but
utilize different dies, possibly to provide faster processors or
more memory. These chips often are used to combine processor and support chipsets so a system-on-chip (SoC) can fit
into a single chip. Intel has taken this approach with a number
of chips, and the company is not alone.
Another way to mix multiple chips together is to stack them
on top of each other (Fig. 6). The dies are not the same size, so
the bonding wires can be attached to each die.
Not all die can be stacked, but the approach has proven
very popular with mobile devices. Smart phones and tablets
often require stacked solutions to address their small size.
Apple’s A4 and A5 SoC solutions for the iPhone and iPad add
a memory die on top of the dual-core processor.
Multichip carriers and stacked die each have challenges.
They are more complex to build and require exacting standards. Power distribution and heat distribution are also significant issues in most applications. The issues depend upon
the chips employed and how they are utilized.
For example, multichip carriers like those used with highend microcontrollers have die that dissipate considerable
amounts of heat. Separating them in a multichip carrier
enables better heat distribution to heatsinks. A stacked die
would be a challenge because the heat dissipation also can
affect how the die are mounted.
Stacked die will generate heat, which must be kept within
a manageable range. That’s why a smart phone can get rather
warm, but it shouldn’t get as hot as a processor chip within a
PC. Two-level stacked die are common. Multilevel stacked die
5. Freescale’s 32-bit Arm Cortex-M0+ Kinetis KL02 wafer scale package is only 1.5 mm by 2 mm.
COMPOUND CHIPS
A package usually includes a single die. This simplifies the
overall design, and it’s normally the configuration required for
a chip. A larger die can address higher-density applications.
Unfortunately, moving to a larger die is not always an
option. In other instances, multiple dies may be required
because each die needs to be created separately. This is often
the case when different technologies are required. For example, some analog or power circuits are better when they are not
built on the same die as the digital components.
There are several ways to create a system-in-package (SiP),
not to be confused with single inline packages (SIPs), which
are simpler DIPs. Multichip SiPs come in a number of forms
GO TO ELECTRONICDESIGN.COM
37
EngineeringEssentials
Die
Die
Circuit board
Bonding wire
Metal layer
Solder balls
7. Xilinx uses its
2.5D multidie tech-
6. Stacked die packages layer additional die on top of each other.
These die are then wired together to provide a cohesive system.
nology in its high-end Virtex-7
Designs need to address power and heat distribution due to mul-
FPGAs. It connects the multiple die using
tiple die.
are possible, but additional levels increase system complexity.
Construction obviously requires more steps.
Multichip packages have other advantages. Each die in a
package can be tested and verified prior to package construction. This makes it more likely that the package will result in
a working system. Likewise, die can be binned (categorized)
based on functionality and performance.
Multichip packages reduce wiring distances. Wiring density, although less than on-chip, normally is higher than off-chip
wiring density. Likewise, the in-package wiring can be more
efficient. It also is not as susceptible to the printed-circuit
board (PCB) design. High-speed serial connections and parasitic capacitance issues tend to exacerbate these issues.
Multichip solutions are becoming more common because
of the variety of technologies needed these days. Nanotechnology is often utilized for sensors like accelerometers. Developers are also looking for smart sensors and matching them
up with a microcontroller. Mixing multiple sensors with the
microcontroller allows sensor integration in a single package.
a metallic interposer layer.
Improving the design tools was paramount in producing
FPGA designs that take advantage of the multislice architecture. Xilinx’s Vivado design tool does this (Fig. 9). It knows
about the multislice nature of the Virtex-7 as well as how the
interconnects link the slices together.
The 2.5D interposer approach is similar to the multichip
carrier except on a much smaller, finer scale. The approach
works very well with FPGAs that need a high-density slice-toslice interconnect. This more regular interconnect approach
could be applicable to mixing other array-style dies such as
memory die including DRAM and flash or other non-volatile
storage. Here the wide parallel connects would be more efficient, eliminating the transceivers and SERDES now found in
many high-density memory interfaces.
The approach would be more challenging for mixed logic
slices since the technology is relatively new. Still, the approach
would be a good way to mix a hard-core processing slice with a
set of FPGA slices and high-speed SERDES. Nothing like this
has been announced yet, and it will be at least a couple of years
before it might occur.
GOING PAST 2D
Multichip solutions result in more dense packages, but
moving toward the die in terms of design can reduce size further while increasing connection density. This is important in
a number of applications.
Xilinx’s 2.5D technology takes this approach (Fig. 7). It utilizes a 65-nm metallic interposer layer to connect up to four dies
or “slices.” The slices can be homogeneous or heterogeneous.
The Virtex-7 H580T employs two 28-nm FPGA slices and
a high-speed SERDES slice with eight 28-Gbit/s transceivers
implemented in 40-nm technology. The SERDES can be
mixed with the FPGA slices, but the result is suboptimal in
terms of performance and efficiency. The hybrid approach
provides the best of both silicon technologies.
The interposer provides a significantly higher sliceto-slice interconnect density that would be possible
using bonding wires that are in turn better than doing
the same thing off-chip. There are tens of thousands
of connections between slices.
38
3D MEMORY STACKS
Stacked die are one form of die-on-die packaging technology that has a long track record. Stacked 3D solutions also are
emerging. They employ through-silicon vias (TSVs) to create
a true 3D chip. They have the advantage in terms of connection density and efficiency, but they do have heat and power
distribution issues like the stacked-die approach. Power has to
be distributed through the die to the next one above. The same
is true for heat.
The 3D stacked chip holds a lot of promise and may
eventually be as common as multilayer PCBs
are now. For now, memory solutions
are taking advantage of this approach.
8. The Xilinx Virtex-7 H580T employs two
28-nm FPGA slices and a high-speed SERDES
slice with eight 28-Gbit/s transceivers implemented in 40-nm technology.
07.11.13 ELECTRONIC DESIGN
(a)
(b)
9. Xilinx’s Vivado design tool accounts for the multislice nature
10. The Hybrid Memory Cube (HMC) stacks multiple memory die on
of the Virtex-7 when turning an initial layout (a) into an optimized
top of a logic layer. The die are stacked on top of each other, and they
layout (b).
are connected via though-silicon vias (TSVs).
The Hybrid Memory Cube (HMC) Consortium has delivered its first specification, which defines an architecture with a
logic layer and multiple, identical memory die stacked on top
of it in a 3D TSV configuration (Fig. 10).
The logic layer provides the access mechanism to the memory die and the high-speed serial interface to the outside
world. The approach eliminates the bonding wires found in
stacked die solutions using TSVs for connections between
memory chips and to the logic layer. An HMC has four or
eight memory die. A board designer will only need to contend
with the capacity and standard interface in a fashion similar to
DRAM modules today.
There are architectural differences between a DRAM module and an HMC. DRAM modules are designed to work with
a single host, although that host may be part of a multicore
NUMA system. HMCs are designed to work with multiple
hosts and can be employed in an array (Fig. 11). An array of
HMCs increases the capacity potential but also the system’s
bandwidth potential. A host attached to any point of the array
has full access to the entire memory.
Go With The Best In Motor Control!
=LORJ0RWRU&RQWURO6ROXWLRQV
%/'8QLYHUVDO
DQG
%UXVKHG'&0RWRU&RQWURO
6WHSSHU
0RWRU
&RQWURO
=(QFRUH)$6HULHV3LQ
=(QFRUH;3)6HULHV3LQ
3KDVH6LQJOH3KDVH
$&,QGXFWLRQDQG3060
0RWRU&RQWURO
3HUPDQHQW0DJQHW6\QFKURQRXV0RWRUV
=)0&6HULHV0RWRU&RQWURO
For more information, please visit www.zilog.com.
GO TO ELECTRONICDESIGN.COM
%/'&6HQVRU
DQG
6HQVRUOHVV0RWRU&RQWURO
=)0&6HULHV0RWRU&RQWURO
Design With Freedom
39
EngineeringEssentials
Multiport support of a single HMC has design implications
for multihost environments such as network switching. It simplifies host memory management and provides scalable support that would otherwise require more complex host designs
like those found in high-end, multicore processors.
The multiport HMC interface uses high-speed serial links
like PCI Express but with its own HMC protocol. It is scalable
by lane like PCI Express, and a single HMC chip can support
hosts with different link requirements. The interface is host
agnostic, although a host will obviously need an HMC interface. This should be a software design exercise for FPGAs.
HMC represents a major change in memory architectures. It
should be able to provide a 15x bandwidth increase compared
to DDR3 while cutting power per bit by 70%. The 3D architecture reduces the
memory footprint
by 90%. Of course, the memory system should have a lower
latency and higher availability than existing memory solutions.
Companies like Micron hope to deliver HMC chips soon
(Fig. 12). Initially, HMC is targeting high-performance applications such as networking and servers. Thicker chips will
stand out , but initially they will be much like packages already
found on systems. Standard pin-outs will make designing with
HMC chips easy with multiple sources for the hardware. HMC
hardware will be available in the next year or so.
HMC targets DRAM. In the future, it could address nonvolatile storage. Designers are mixing flash and DRAM like
Viking Technology’s ArxCis-NV hybrid DDR3 DRAM module, which also incorporates
flash backup storage. It
might simply be a matter of
adding flash memory die to
the HMC stack.
11. HMCs can be connected in an array to
12. The HMC package is
increase density and
designed to deliver a 15x per-
t h ro u g h p u t w i t h a
formance improvement over
number of hosts.
existing DDR3 modules while
using 70% less energy.
&
The Convenient All-in-One Solution
for Custom-Designed Front Panels & Enclosures
present
FREE
Software
Embedded RFID
• Authenticating
• Tracking
• Protecting
ONLY $90.24
with custom
logo engraving
vailable
ction
NOW a ntenna conne
ha
Get your UHF RFID Kit
at www.beta-eSTORE.com
www.magic-pcb.com
www.pcb-pool.com
40
PCB-POOL® is a registered trademark of Beta LAYOUT GmbH
- wit
tities
ll quan
- for a
You design it
We machine it
to your specifications using
our FREE CAD software,
Front Panel Designer
and ship to you a
professionally finished product,
no minimum quantity required
ost effective prototypes and production
C
runs with no setup charges
OPowder-coated and anodized finishes in
various colors
O
O
Select from aluminum, acrylic or provide
your own material
O
Standard lead time in 5 days or express
manufacturing in 3 or 1 days
FrontPanelExpress.com
1(800)FPE-9060
07.11.13 ELECTRONIC DESIGN
Q&A
BILL WONG | Embedded/Systems/Software Editor
bill.wong@penton.com
Matterform Founders
Discuss Their 3D Scanner
E
ven the mainstream media has been covering 3D printers and how they are
going to revolutionize manufacturing, especially for small businesses (see
“Tap New Technologies To Produce Practically Perfect Prototypes” at electronicdesign.com). So if 2D printers have 2D scanners, why shouldn’t 3D printers have 3D
scanners? The Matterform 3D Scanner costs less than $600 (see the figure). Company
cofounders Adam Brandejs and Drew Cox explain the technology.
WONG: Tell me about the scanner.
MATTERFORM: It’s the world’s first
truly affordable 3D scanner for anyone!
It allows you to take a physical object
and turn it into a digital 3D model on
your computer. From there, you can
print your file on any 3D printer or
online printing service or use the model
you created in an animation or video
game. It’s lightweight, portable, and
compact, making it easy to integrate
into your workspace. Anyone can use it.
End of story. Unfold it, plug it in, place
an object on the scan bed, and press go!
It’s that simple. Really.
Not only is it the first affordable 3D
scanner, but for the first time ever in the
home 3D makerspace comes a product
that you can actually be proud to have
on your desk. Long gone are the days of
laser-cut wood and threaded rod projects. Get creating and scanning right out
of the box. When not in use, the scanner
easily folds and tucks away on a shelf,
and because of the built-in handle it also
makes it easier for you to transport it
should you ever want to.
Its high-resolution scan is created with
a high-definition camera and dual lasers.
All scanners are fully assembled and
tested prior to shipping. There are no kits
here. Its easy folding design makes storage or transport simple. Multiple quality
GO TO ELECTRONICDESIGN.COM
settings can be chosen, but an average
scan is 3 minutes. Scan objects up to
190 mm by 190 mm by 250 mm (7.5-in.
diameter and 9.75-in. height). Dual stepper motors provide full software control
over the scan bed and the z-axis. See the
point cloud being captured in real time!
WONG: How does it work?
MATTERFORM: The unit works from
a standard 5-V dc power supply and
is connected to the PC via a single
USB 2.0 connector. The software was
designed from the ground up and works
seamlessly with the hardware. Free to
use and download, the software is available for PC, Mac, and Linux. Whether
you’ve used 3D software before or if
you’re just getting into it now, you’ll
find the software was designed to be as
easy as possible. After a one-time setup,
scanning becomes as easy as one click.
We know you want to spend your time
creating things and being creative, not
calibrating, tweaking hardware, or figuring out yet another tool.
Because the scanner is one solid
desktop device, the objects you scan are
accurate and precise. Unlike softwarebased solutions, it gives dimensional
data, making 3D scans far more useful.
We’re aiming to make the 3D files compatible with all major CAD packages,
3D modeling software, and 3D printers.
Currently, models produced from the
The low-cost Matterform 3D scanner is the ideal complement for 3D printers.
41
Q&A
over $400,000 contributed so far.
scanner can be saved as .STL, .OBJ, and
point cloud .PLY formats, making it
easy to integrate scans within existing
systems, such as 3ds Max, Maya, SolidWorks, Cinema 4d, Google SketchUp,
Rhino, and TrueSpace.
Currently the resolution of the
scannner, on a 4-in. figurine, is 0.43 mm
at 0.5° scans, with an accuracy of ±0.2
mm. We’re quite excited about the current results and are working every day to
continue improving the resolution.
ing things since he could move his hands.
Professionally, he has been a programmer,
WONG: What future enhancements
artist, art director, and graphic designer. Un-
might be in the works?
MATTERFORM: Many! Including
color.
professionally, he is an inventor, hacker, and
mad scientist. His career path has led him
to help build brands like Labatt, Budweiser,
Audi, Honda, and Coca Cola. His hobbies
ADAM BRANDEJS (left in the photo) is
and passion have brought him to respect
a programmer, sculptor, Eagle PCB junky,
the names of Prusa, Sherline, Arduino, and
hacker, and maker. He has been making
every person who has ever contributed to
things all of his life, from robots and code to
open source projects.
sculptural art that tours art galleries all over
MORE INTERVIEWS ONLINE
the world. He was dry walling at 7, picked
WONG: How does the scanner com-
up his first soldering iron at 10, was pro-
pare to other alternatives?
MATTERFORM: Other alternatives
start at $3000, so the Matterform 3D
Scanner is by far the cheapest.
gramming Commodore 64s at 12, picked
Check out Electronic Design’s series on online
interviews as we probe big industry trends and
particular design issues alike with engineers,
executives, and academics from across the spectrum. For more, go to http://electronicdesign.
com/community/interviews and see:
• LSI Fellow Rob Ober Discusses PCIe Flash
Technology
• Gumstix Explains How To Cut Time-To-Market
Using Modules
• Digia’s Lars Knoll Discusses The Qt GUI
Framework
• Spansion’s CTO Talks About Embedded Charge
Trap NOR Flash Technology
up 3DS Max r2.5 in high school, fixed everyone’s computer during university, and
spent a few years in advertising making
Web sites for Audi, Coca Cola, Budweiser,
WONG: You used crowd funding to get
and countless other big name companies.
started. How has that worked out?
MATTERFORM: Very well. We’ve
achieved over five times our goal with
an entrepreneur since he was 18 and mak-
DREW COX (right in the photo) has been
3:0&RQWUROOHU
+9$PSOLILHU
8OWUDVRXQG,&V
,&V
$UUD\,&V
7HOHFRPPXQLFDWLRQ
3RZHU
/(''ULYHU,&V
0DQDJHPHQW,&V
026)(7V
,&V
+9,QWHUIDFH,&V
Supertex inc.
6SHFLDOW\,&V
Ɣ(OHFWURQLF/LQH6ZLWFK,&V
'HYLFH
6ZLWFK5HVLVWDQFH
,QSXW9ROWDJH
%UHDNGRZQ9ROWDJH
'LDO3XOVH'3'3
+7
ȍPD[
9
9PD[
ȍPD[
9
9PD[
+7
+7
3DFNDJH
$FWLYH+LJK
62,&
$FWLYH/RZ
Ɣ5LQJ*HQHUDWRU,&V
'HYLFH
5LQJ6LJQDO
506
+9
9PD[
PD[
+9
9PD[
RUPD[
+9
9PD[
PD[
2XWSXW/RDG
5(1
+LJK9ROWDJH
3RVLWLYH
6XSSOLHV
+LJK9ROWDJH
1HJDWLYH
6XSSOLHV
([WHUQDO
&LUFXLWU\
<HV
<HV
<HV
1R
1R
6HOIFRQWDLQHG
3DFNDJH
62,&
62,&
/4)3
Ɣ+RWVZDS,QUXVK&XUUHQW/LPLWHU&RQWUROOHU,&V1HJDWLYH6XSSO\5DLO
'HYLFH
+9
+9
,QSXW
9ROWDJH
897KUHVKROG
9ROWDJH
9PD[
9W\S
9W\S
6XSSO\
&XUUHQW
P$W\S
)HDWXUHV
3DFNDJH
8QGHUYROWDJH89GHWHFWLRQFLUFXLWVSRZHURQ
UHVHW325VXSHUYLVRU\FLUFXLWVLQUXVKFXUUHQW
OLPLWLQJVKRUWFLUFXLWSURWHFWLRQDXWRUHWU\
627
Ɣ+LJK9ROWDJH/HYHO7UDQVODWRU,&V
'HYLFH
,VRODWHG2XWSXW
'ULYHUV
+7
'XDO
+7
6LQJOH
,QSXW
9ROWDJH
9
2XWSXW
9ROWDJH
9
9
,QSXWWR
2XWSXW,VRODWLRQ
9
2XWSXWWR
2XWSXW,VRODWLRQ
3DFNDJH
9
')162,&
62,&
7RGRZQORDGGDWDVKHHWVYLVLWZZZVXSHUWH[FRPDQGVHDUFKWKHGHYLFHQXPEHU
42
07.11.13
MEDICAL DESIGN
DesignSolution
LAWRENCE VIVOLO | Synopsys Inc.
lvivolo@synopsys.com
Transaction-Based Verification
And Emulation Combine
For Multi-Megahertz Verification Performance
Transaction-based verification is becoming the emulation mode of choice because of
its unique ability to keep pace with the growing verification support requirements of the
latest high-performance, industry-standard interfaces.
D
SIMULATION
Today, simulation continues to be
the workhorse of complex verification.
Most verification is done using simulation, finding more than 95% of bugs.
GO TO ELECTRONICDESIGN.COM
peripherals
Transactor
Transaction-based
verification (TBV)
SCEMI/ZEMI
Transactor
Simulation usually runs on a single workstation, and its enviesign complexity has grown with each sucronment consists of two components: the design under test
cessive generation of system-on-chip (SoC)
(DUT) and the stimuli generator—the testbench (Fig. 1).
evolution. SoCs now include many indusWith simulation, the design runs at a wire level and is clock
try-standard interfaces, multiple processor
accurate, which means there is a wire-level interface between
cores, on-board memories, and embedded firmware and
the testbench and the DUT. For a full chip, these wires may
software. With this growth, we’ve seen verification complexbe the device pins. With the addition of various monitors and
ity grow exponentially, driving engineers to seek advanced
checkers, the interface between the DUT and testbench can
verification methodologies.
grow to include thousands of wires.
While in-circuit emulation (ICE) promises megahertz veriSimulation is still what’s used for most register transfer level
fication performance, its dependence on protocol-specific
rate adapters combined with the sheer number of high-speed, (RTL) development and design. Nevertheless, for designs
protocol-specific physical interfaces has proven to be a recent requiring hardware/software co-verification, as well as those
challenge for today’s advanced designs.
External hardware
Target system
Host workstation
Transaction-based verification (TBV)
holds the key to bridging the perforTestmance of emulation with the flexibility
DUT
bench
Simulation
and simplicity of simulation.
Simulator
By adopting accelerated transactionbased emulation strategies, designers
TestDUT
In-circuit
bench
can move their verification strategy up a
emulation (ICE)
Emulator
Rate adapter
level of abstraction and achieve the leap
forward in verification performance
Testand productivity that is necessary to
DUT
Embedded
bench
testbench
fully debug and develop the most comEmulator
plex electronic hardware and softwarebased systems. Different use modes are
TestDUT
Co-simulation
available in today’s emulation systems,
bench
(signal level)
Emulator
and TBV is becoming the preferred
emulation mode for achieving multimegahertz verification performance.
Virtual
DUT
Emulator
1. This graphical description of the most commonly used emulation modes highlights the location of the testbench, the design under test, any rate adapters and
physical interfaces, and potential performance bottlenecks.
43
DesignSolution
utilizing many industry-standard protocol interfaces, maximum verification performance is required. Simulation performance is often limited by the performance of the host PC
or workstation on which it executes. A large SoC, designed
to run at 1 GHz, may clock at 1 Hz or less in simulation on a
standard PC.
To put that into perspective, if the operating-system (OS)
boot-up sequence for a cell phone, whose processor is running at 1 GHz, were to take 10 seconds, the same boot-up
sequence, simulated on a host PC at 1 Hz, would require 1
billion seconds (317 years). Very few design teams enjoy a
project schedule that can accommodate that.
As a result, more and more design teams are adopting some
form of hardware-assisted verification. Today’s high-performance emulators run very fast and support many operating
modes. Synopsys’ Zebu, for example, reduces the boot-up
time for the above example from more than 300 years to less
than an hour with its transaction-based verification mode.
IN-CIRCUIT EMULATION
The first widely used emulation platforms were introduced
in the early 1990s and were limited to ICE. With ICE, the idea is
to eliminate the host PC or workstation, which was (and still is)
the performance bottleneck. First, the DUT is moved from the
host PC to dedicated emulation hardware. Next, a live “target
system,” or a physical device, replaces the PC-based testbench.
A physical cable connects the two. The cable is often
plugged into the very socket that the SoC will eventually plug
into. This combination eliminates the host workstation or PC,
allowing the design to run several orders of magnitude faster,
typically at the maximum performance of the emulator itself.
ICE is not without its challenges, however, as an emulator is
rarely fast enough to run at the same speed at which a target
system would normally operate. In the early days, the entire
target system had to be slowed down to match the speed of
the emulator. This tended to be very complicated, and many
devices, such as disk drives, simply could not be slowed
down. Thus, rate adapters were introduced.
A rate adapter acts as a buffer, collecting all the required
information from the DUT and—once the data (wire-level
signal changes) is sufficiently cached—sending it at real-time
speed to the target system. It also captures data from the
target system at full speed, caches it, and then plays it back at
emulation speed to the DUT.
While this approach works well for fairly simple protocols,
today’s SoCs include many complex, high-speed interfaces,
each of which would require a rate adapter to be connected
to an emulator. Even when a rate adapter is available, the
constant evolution of speed and complexity of individual
protocols (such as SATA2 versus SATA3) has made timely
rate adapter development difficult. Speed requirements alone
44
typically delay, and in more recent cases eliminate, the availability of rate adapters for the latest protocols.
Another characteristic of ICE is the simple fact that the
physical devices connected to the emulator often require an
operator to be physically present—for example, to push a
reset switch or power-cycle the target system. This makes it
difficult to leverage an emulator that is centrally located.
EMBEDDED TESTBENCH
In an attempt to leverage the existing simulation environment, as well as to avoid some of the limitations of ICE, many
design teams in the mid-1990s adopted an embedded testbench approach that moves the entire simulation testbench
onto the emulator. This eliminates the physical interfaces
required by ICE, and with them, the need for rate adapters.
Unfortunately, such a testbench must be synthesizable to
be loaded onto an emulator. Because virtually no simulation
testbenches are synthesizable in today’s advanced verification
environments, this method is limited in use.
A variation (special case) of this approach is commonly
used today. Leveraging the emulator’s large reserves of physical memory, designers load this memory with system software, such as firmware, drivers, or an entire OS. The DUT’s
embedded processor can then run that software directly as if
it were running in the actual system, effectively making the
software a testbench. This approach is relatively simple to
implement, and since there are no physical connections, the
emulator runs at its highest level of performance
As the software is only being run through the processor,
this approach still lacks the ability to test interfaces between
different chip protocols. Additionally, because it requires a
fairly complete processor implementation, it can only be used
relatively late in the design cycle. As a result, it is generally
best suited for late system-level hardware verification and
early hardware/software bring-up.
CO-SIMULATION
With co-simulation, also called simulation acceleration, the
DUT is moved to the high-performance emulator as with the
other use models. The testbench, though, remains on the host
PC or workstation. A wire-level interface connects the two.
This verification approach is very straightforward, as it
leverages the existing simulation testbench and eliminates the
need for external rate adapters. However, the actual verification performance is heavily limited by the performance of the
host PC, the complexity of the existing testbench, and/or the
wire-level interface between the testbench and the DUT.
In a typical verification environment, there are thousands
of signals at the interface, with many changing multiple times
within each clock period. With wire-level co-simulation, each
signal must be transferred between the testbench and DUT on
07.11.13
ELECTRONIC DESIGN
Emulator
PC
magnitude faster than any wire- or pinbased verification approach.
Front-end
TX
Back-end
Testbench
Communication
C/C++/
Beyond performance, TBV offers
DUT
HDL BFM
C/C++/
infrastructure
SC/SV
RX
model
designers
several other advantages.
SC/SV
model
Primarily, it eliminates the need for rate
adapters and physical interfaces. With
TBV commands
Bit-based protocol
TBV, each physical interface is replaced
with a transaction-level interface.
2. In the transaction-based verification environment, the back-end portion of the
transactor and the DUT, which remains unchanged, are located within the emulator. Likewise, the rate adapter, required for
ICE, is replaced with a protocol-specific
This requires that they both be written with synthesizable RTL.
transactor.
Unlike
rate
adapters,
transactor models for the latest protoevery transition. Even though the emulator can run orders of
cols
are
readily
available
off-the-shelf and are easily upgraded
magnitude faster, it must wait for these transfers to complete.
to
accommodate
protocol
revisions. Synopsys today, for
In addition, the emulator will often be stalled by the perexample,
provides
vast
libraries
of transactors for standard
formance of the testbench itself. If tightly coupled to the testinterface
protocols
as
well
as
tools
to enable the development
bench, which is common in pure simulation environments,
of
custom,
proprietary
transactors.
the emulator will quickly stall, waiting for the slower testIt is also possible to create an emulation-like environment
bench to react to incoming signals and produce the next set of
by
using transactors to connect the DUT to “virtual devices.”
stimuli. As a result, most designers do not use this mode for
A
virtual
device is simply a soft model of a device that runs
general verification. But they do use it as a stepping stone (to
on
the
host
PC. A design using an HDMI transactor to test an
validate the DUT functionality on the emulator) for TBV.
HDMI interface, for example, would be able to leverage the
same transactor to connect to an HDMI virtual display. This
TRANSACTION-BASED VERIFICATION
would allow the engineer to see the video output from the
TBV raises the level of verification abstraction away from a
wire-level interface to run thousands to a million times faster
than simulation on a host PC. It simplifies the communication between the testbench and DUT so a design team can
access the full performance benefit of the emulator.
With TBV, like co-simulation, the DUT is loaded onto the
emulator, and the testbench resides on a host PC. But instead
of a wire-level interface, TBV uses a high-performance interface using transaction-level protocols that are executed on the
emulator. This transaction-level interface allows the testbench
to be tightly or loosely coupled to the DUT.
The communication between the testbench, now working at
a protocol level, and the DUT, which still requires a wire-level
interface, is accomplished through a transactor. The transactor
converts the high-level commands from the testbench into the
wire-level, protocol-specific sequences required by the DUT,
and vice versa. The key is that all the wire-level communications are wholly contained within the emulator itself and run
orders of magnitude faster as a result (Fig. 2).
Another key difference of transaction-based verification
is that the transaction-level interface allows the testbench to
stream data to the DUT, which the transactor buffers automatically. This speeds up the execution of the testbench. With
this methodology, it is also possible to have multiple trans5R+6
actions active across multiple transactors. Together, these
transactors enable the emulator to process data continuously,
which dramatically increases overall performance to that of
a pure ICE environment. As noted earlier, this is orders of
Transactor
9
GO TO ELECTRONICDESIGN.COM
45
DesignSolution
DUT just as if it were connected to a physical monitor, without the need to physically connect to an actual monitor.
Some of today’s virtual devices are bridges that provide
connectivity to the actual physical device, but via the host
PC (and a transactor). A USB bridge, for instance, connects
a physical USB device, such as a memory stick, to the USB
port of the host PC. As designed, the DUT only sees the USB
46
drive and behaves accordingly once it detects the presence
of the flash drive (for example, it installs a driver, if the OS is
also being emulated).
An additional distinction of TBV is remote accessibility. As
there are no physical interfaces connected to the emulator, a
user can fully use and manage the emulator from anywhere
in the world. When combined with virtual devices such as
video displays, protocol analyzers, and
audio processors, this distinction is
very obvious.
A user from anywhere in the world
can remotely view a virtual display.
An ICE user, on the other hand, would
need to be in the same room with the
TV, which is cabled directly into the
emulator. Remote accessibility is also
available when connecting the emulator to an architectural model environment, such as Synopsys’ Platform
Architect and/or Virtualizer, which
would also be done through a transaction-level interface.
A transactor consists of three parts:
the front-end interface, a back-end RTL
bus-functional model (BFM), and a
physical communications channel.
The front-end interface is typically
a behavioral model that runs on the
host PC and interfaces to the testbench,
usually through Direct Programming
Interface (DPI) calls. It is written in C/
C++, SystemC, or System Verilog. This
piece converts high-level transactionlevel commands and sends them across
the physical interface protocol using
either the Standard Co-Emulation
Modeling Interface (SCEMI) standard
or the Zebu Emulation Modeling Interface (ZEMI) language.
The interface channel between the
front-end interface on the host PC and
the BFM on the emulator may be implemented at the physical level using any
high-performance interface standard
such as PCI Express. A standard TBV
protocol, such as SCEMI or ZEMI, runs
on top of this physical interface. This
protocol is transparent to the user.
The back-end hardware description
level (HDL) BFM runs on the emulator
and interfaces with the communica-
07.11.13
ELECTRONIC DESIGN
tions channel to send and receive transactions. Here, the
transactions are converted to a wire-level interface that
interacts with the DUT. Because the BFM resides on the
emulator, it must be fully synthesizable. The BFM also must
adhere to the target interface protocol and therefore must be
fully functional.
So, as an example, the BFM would receive the high-level
transaction command “read file” from
the channel and convert it to the precise,
cycle-accurate, wire-level handshake
sequences that a physical direct memory
access (DMA) controller, located within
the emulated DUT, would need.
tion can be used throughout the verification cycle, with
worldwide 24x7 access.
LAWRENCE VIVOLO is the product marketing director,
emulation, at Synopsys. He holds a BS in engineering from
California Polytechnic State University, San Luis Obispo, and
an MBA from Santa Clara University, Santa Clara, Calif.
TBV IN THE DESIGN FLOW
TBV can be used throughout the flow,
from the block level and beyond. Common applications include:
• Debug and verification of large blocks,
subsystems, or complete SoCs
• Driver development or system-level
verification when combined with virtualization
• Early hardware/software bring-up,
including firmware, drivers, and OSs
• Full-chip power analysis and estimation
• System-level hardware/software coverification or early software development when connected to prototyping
hardware
CONCLUSION
TBV with emulation offers design
teams a unique opportunity to accelerate
SoC verification. By raising the level of
abstraction of their verification environment, teams can achieve multiple orders
of magnitude improvement in their verification performance. With TBV, designers have access to a platform that delivers
megahertz performance for block- and
system-level verification, as well as early
hardware/software bring-up.
The use of transactors delivers all the
benefits of in-circuit emulation without
the challenges of rate-adapter availability
and physical accessibility. When combined with virtual interfaces and systemlevel virtual models, TBV with emula-
GO TO ELECTRONICDESIGN.COM
47
ideas
for design
Graphical, Numerical Techniques
Size LED Array’s Drive
KENG WU | SWITCHING POWER INC. kengchi.goah@gmail.com
L E D - BA S E D L I G H T I N G I S R E P L AC I N G
incandescent and gas-discharge lamps
in many situations. As practical lighting
sources, and considering their directional nature, LED fixtures usually use
multiple LEDs arranged in an array,
with “m” LEDs in a serial string and “n”
such strings in parallel (Fig. 1). For the
array configuration and the exponential I-V characteristics model of a single
LED,1 it’s additionally possible to have a
similar model representing the loading
of such an array.
Since the single LED is a diode with
a nonlinear forward-voltage/current,
driving it with a fixed voltage source is
generally not recommended. Rather,
a serial (current-limiting) resistor is
needed to equalize the power source and
the load (Fig. 2). However, sizing this
resistor requires a significant analytical manipulation. Using the single-LED
exponential model, a Kirchoff ’s voltage
law (KVL) equation around the circuit
loop gives:
separating the two key I-V relationship
equations, v + RxIF = Vs and IF = aebv + c.
For example, using a Lumileds Luxeon Rebel ES LED with a 3.2-V dc source
and a desired drive current of 0.5 A, the
horizontal current line intercepts the
LED curve at 2.92 V (Fig. 3). Therefore,
the series resistor will be R x = (3.2 –
2.92)/0.5 = 0.56 Ω.
However, the graphical approach
would not handle an LED array effectively, and the numerical method works
better. To represent the array current iA,
the exponential model for the (m × n)
array is modified, since voltage drops
across current-limiting resistor Rn and
sense-resistor R Sense are usually small
compared with driving-source Vs:
9 ²9
5 [ VE9) )
DH F
R1
D1
48
9V
P
F
With the array model, the switchmode buck-regulating current driver
in a closed-loop configuration (Fig. 4)
can be described by a set of equations
starting at the feedback point, circling
around the control loop, and ending at
the driver output:
• Current sensing:
Vf = iA • RSense
(4)
• Error amplifier:
Ver = A(VRef – Vf )
(5)
9HI 9HU ² 9)
% &75 % 5 H 5G
• PWM ramp:
95DPS W 90LQ 90D[ ² 90LQ
W % 7V
Dm
Rx
where a, b, and c are model parameters
for a selected LED, as defined in the
reference.
This is a challenge to solve symbolically, but it can be solved numerically
with mathematical software tools. Alternatively, it can be solved graphically by
E
• Effective error:
Rn
5 [ DH E9) F 9) 9V which can be rearranged to:
L$ Q DH
VS
IF
VF
R Sense
1. Establishing the correct resistor size for
2. Even for a single LED driven by a voltage
each string of an array of n strings, where
source, the exponential current-voltage
each strings has m LEDs, is challenging
relationship complicates selection of the
due to the nonlinear nature of the diodes.
series-resistor value.
07.11.13 ELECTRONIC DESIGN
1μA IQ Synchronous Boost Converter Extends Battery Life in
Portable Devices
Design Note 516
Goran Perica
Introduction
Boost converters are regularly used in portable devices to produce higher output voltages from lower
battery input voltages. Common battery configurations include two to three alkaline or NiMH cells or,
increasingly, Li-Ion batteries, yielding a typical input
voltage between 1.8V and 4.8V.
Because the batteries used in portable devices are
usually as small as possible, they present high internal impedance under heavy loads, especially close to
the end of their discharge cycle. Unlike other boost
converters that struggle with high source impedance
at startup, the LTC3122 prevents high surge currents
at startup.
The 12V output converter shown in Figure 1 is designed
to run from any typical small battery power source. This
design centers around the LTC®3122 boost converter,
which can efficiently generate a regulated output
up to 15V from a 1.8V to 5.5V input. The LTC3122
includes a 2.5A internal switch current limit and a full
complement of features to handle demanding boost
applications, including switching frequency programming, undervoltage lockout, Burst Mode® operation or
continuous switching mode, and true output disconnect.
The integrated synchronous rectifier is turned off when
the inductor current approaches zero, preventing reverse
inductor current and minimizing power loss at light loads.
1.8V to 5.5V Input to 12V Output Boost Regulator
The circuit in Figure 1 is designed for high efficiency
and small size. The LTC3122 operates at 1MHz to
minimize the size of the filter capacitors and boost
inductor, and uses Burst Mode operation to maintain
high efficiency at light loads, as shown in Figure 2. At
heavier loads, the converter can operate in constant
frequency mode, resulting in lower input and output
ripple. Constant frequency operation can result in
lower EMI and is easier to filter.
This unique output disconnect feature is especially
important in applications that have long periods of idle
time. While idling, the part can be shut down, leaving
the output capacitor fully charged and standing by for
quick turn-on. In shutdown, the part draws less than
1μA from the input source.
70
CAP
SW
0.1μF
VOUT
PGND
PWM/SYNC
SD
VIN
RT
PGND
VC
VOUT
12V
R1
1.02M
FB
VCC
RT
57.6k
COUT
0.1μF
CAP
VIN
CVCC
4.7μ
RC
R2
86.6k 113k
CC
330p
dn4gp F01
Figure 1. The 1MHz Operating Frequency and Small
Inductor Make This Converter Suitable for Demanding
Portable Battery-Powered Applications.
90
80
U1
LTC3122
4.7μH
XAL4030-472
CIN
10μF
Efficiency can be improved further by increasing the
inductor size. Figure 4 shows the increase in efficiency
EFFICIENCY (%)
VIN
1.8V to
5.5V
Efficiency can be increased by running the LTC3122 at
a relatively low switching frequency. Figure 3 shows
the results of reducing the switching frequency from
1MHz to 500kHz.
60
50
40
30
20
VIN=2V, BURST MODE
VIN=3.3V, BURST MODE
VIN=2V, PWM MODE
VIN=3.3V, PWM MODE
10
0
0.1
1
10
IOUT (mA)
100
FSW = 1MHz
1000
dn4gpf F02
Figure 2. The High Efficiency of the LTC3122 Boost
Converter Extends Battery Life in Portable Applications.
L, LT, LTC, LTM, Burst Mode, Linear Technology and the Linear logo are
registered trademarks of Linear Technology Corporation. All other trademarks
are the property of their respective owners.
achieved by replacing the 4mm w 4mm boost inductor
(XAL4030-472) with a 7mm w 7mm inductor (744-777910 from Würth). The 90% efficiency at 10mA is 5%
higher than the efficiency shown in Figure 3.
90
80
EFFICIENCY (%)
70
60
50
40
30
1MHz, BURST MODE
500kHz, BURST MODE
1MHz, PWM MODE
500kHz, PWM MODE
20
10
0
0.1
1
10
IOUT (mA)
100
1000
dn4gpf F03
Figure 3. The Efficiency is Greatly Affected by the Operating
Frequency. At 100mA Load an Additional 4% Can Be Gained
by Reducing the Switching Frequency from 1MHz to 500kHz.
source can lift the output rail to regulation. The input
current slowly ramps up. The input current overshoot
required to charge the output capacitor is limited to
only 200mA and the input power source voltage droop
is limited to 0.5V, as shown in Figure 5.
Conclusion
The LTC3122 boost converter serves the needs of
battery-operated applications that require low standby
quiescent current and high efficiency. Unlike many
other boost converters, it includes features, enableing operation from batteries near full discharge when
battery ESR becomes high. Its very low quiescent and
shutdown currents, combined with output disconnect,
extend battery run time in applications with long idle
periods. The LTC3122 includes a complete set of
features for high performance battery operated applications and comes in a small, thermally enhanced
3mm w 4mm package.
90
Output Disconnect
Typical boost converters cannot disconnect the output
from the input because of the boost diode. Current
always flows from the input through the inductor and
boost diode to the output. Therefore the output can not
be shorted or disconnected from the input, a significant
problem in many applications, especially in shutdown.
In contrast, the LTC3122 includes an internal switch
that disconnects the boost MOSFET body diode from
the output. This also allows for inrush current limiting
at turn-on, minimizing the surge currents seen by the
input power source.
80
70
EFFICIENCY (%)
Battery size should be taken into account when considering inductor size. Using a relatively small inductor
running at a high frequency may necessitate a correspondingly higher capacity battery to achieve the
same run time at relatively lower efficiency. In other
words, space gains achieved with a smaller inductor
may be replaced by the need for a bigger battery.
60
50
40
30
VIN = 2V, BURST MODE
VIN = 3.3V, BURST MODE
VIN = 2V, PWM MODE
VIN = 3.3V, PWM MODE
20
10
0
0.1
1
10
IOUT (mA)
100
1000
FSW = 300kHz
dn4gpf F04
Figure 4. With a Lower Switching Frequency and a Larger
Inductor, a Smaller Battery Can Be Used. Efficiency Gain
Up to 30% in the 1mA to 10mA Load Range (in PWM Mode)
Can Significantly Improve Applications That Operate with
Light Loads.
Figure 5 shows the output of the LTC3122 disconnected
in shutdown. The output voltage is pulled to zero by the
load following shutdown, and the LTC3122 consumes
less than 1μA of current.
Start-Up Inrush Current Limiting
To simulate a real battery-operated application, the
circuit in Figure 1 was tested with 1Ω of equivalent
series resistance (ESR) placed between the power
source and the LTC3122 circuit. Once the LTC3122 is
enabled, it controls the startup so that the input power
Data Sheet Download
www.linear.com/LT3122
dn4gpf F05
Figure 5. Inrush Current Limiting at Turn-On Minimizes
Surge Currents Seen by the Input Source. The Output Is
Disconnected from Input During Shutdown.
For applications help,
call (408) 432-1900, Ext. 3788
/LQHDU7HFKQRORJ\ &RUSRUDWLRQ
GQI/7$3.‡35,17(',17+(86$
FAX: (408) 434-0507 O www.linear.com
¦/,1($57(&+12/2*<&25325$7,21
1630 McCarthy Blvd., Milpitas, CA 95035-7417
(408) 432-1900
O
*'+
*'*
*')
)'2
)'1
• Duty cycle, D:
(8)
<nkk^gm!:"
VRamp(D • Ts) = Vef
• Power stage:
9,Q
% 1 V % ' 9V
1S
b? ! o " )'0
)'/
Bk ! o "
)'.
)')',
)'+
With effort, all the equations can be consolidated into a
single, closed-loop equation. Equation 10 is a transcendental
equation with an exponential term that prevents us from solving it analytically, but it can be solved using computational
techniques and software tools.
)'*
)
+'/
+'0
+'1
+'2
,')
,'*
,'+
OhemZ`^!O"
3. The intersection of the desired current line and the forward cur-
Reference
rent/voltage curve (red) leads to the unique solution to the non-
1. “Generate Realistic Models for LED Current Versus Voltage,” Keng C. Wu;
Electronic Design, Vol. 61, No. 1, p. 52, Jan. 10, 2013, http://electronicdesign.com/power/generate-realistic-models-led-current-versus-voltage
linear equations of the LED model. (The blue line is the reverse-bias
current/voltage relationship.)
9V
$ 95H I ² Q DH
E
KENG C. WU has a BS from
& % 5 6HQVH ² 9)
P
5G
9,Q
% 1V %
1S
Chiaotung University, Taiwan,
% &755H²90LQ
University, Evanston, Ill. He has
published four books and holds
9V
90D[ ² 90LQ
VIn
and an MS from Northwestern
seven U.S. patents.
VS
D1
Np
NS
D2
Figure 1
array
Ver
VMax
Driver
VMin
PWM
TS
–
A
+
Rd
VCC
Vf
VRef
–
+
CTR
VF
RSense
Vef
Re
4. Analysis of the switch-mode buck-regulating current driver in a closed-loop configuration for driving an array yields a set of equations
that are better solved using numerical computation rather than an analytical, closed-form approach.
GO TO ELECTRONICDESIGN.COM
GO
49
:JHU/LYL
$6$
TV\ZLYJVT
;OL5L^LZ[7YVK\J[ZMVY@V\Y5L^LZ[+LZPNUZ
&HUWLILHG'LVWULEXWRU
>LKLSP]LY
(:(77+8
HUK10;:V
`V\»YLUL]LY
:63
4V\ZLYKLSP]LYZ[OLJVTWVULU[Z`V\ULLKVU[PTL(UK^P[OSVJHS;LJOUPJHS
:\WWVY[HUK*\Z[VTLY:LY]PJL,_WLY[ZPU SVJH[PVUZHYV\UK[OL^VYSK`V\»SS
ÄUK[OLUL^LZ[JVTWVULU[Z[VSH\UJO`V\YUL^KLZPNUZLHTSLZZS`
(\[OVYPaLKKPZ[YPI\[VYVMZLTPJVUK\J[VYZ
HUKLSLJ[YVUPJJVTWVULU[ZMVYKLZPNULUNPULLYZ
4V\ZLYHUK4V\ZLY,SLJ[YVUPJZHYLYLNPZ[LYLK[YHKLTHYRZVM4V\ZLY,SLJ[YVUPJZ0UJ6[OLYWYVK\J[ZSVNVZHUKJVTWHU`UHTLZTLU[PVULKOLYLPUTH`IL[YHKLTHYRZVM[OLPYYLZWLJ[P]LV^ULYZ
from
globalpurchasing.com
VICTORIA FRAZA KICKHAM | DISTRIBUTION EDITOR | victoria.kickham@penton.com
JULY 11, 2013
Globalization,
Product Obsolescence
Top Supply Chain
Challenges
Distributors respond with tools to
help engineers and procurement
professionals meet the increasingly complex electronics supply
chain, says Components Direct’s
Steve Martin.
DISTRIBUTION RESOURCE
RECENTLY CAUGHT up with
Slow Recovery
Still The Norm
Electronics industry executives adjust to slower
industry conditions as they look to new regions and
expanding markets for electronic content as key
growth drivers.
SPEAKING TO OTHER ELECTRONICS
supply chain executives in May, TTI’s
Michael Knight characterized 2012 as a year in which the electronics
industry “went sideways,” explaining the flat to very low growth most distributors in the sector experienced compared to 2011. This is new for an
industry accustomed to sizeable year-over-year sales gains driven by new
technologies and an ever-growing need for electronic content in just about
all aspects of daily life.
“I think there’s a possibility that 2013 will be another ‘sideways,’” said
Knight, TTI’s vice president, Americas, in a separate interview in early
May. “I think we’ll see some slight ‘up,’ but that’s not how the industry historically works. Plus or minus 2% or 3%, that’s just not how our industry
looks. But it’s happening.”
The slow economic recovery globally, combined with political uncertainty and other pressures here at home, is the main culprit, and most distribuContinued on Page 52
GO TO ELECTRONICDESIGN.COM
GO
Steve Martin, executive
vice president of sales at
Components Direct, a
franchised distributor
that specializes in excess,
obsolete, and end-of-life
electronic components.
Steve Martin
We asked Martin to discuss the state of the electronics supply chain and
provide an industry outlook for the remainder
of the year. He pointed to globalization, product
obsolescence, and the increasingly important
role of electronic commerce as key issues shaping the electronics supply chain. What follows
are excerpts from our conversation.
DISTRIBUTION RESOURCE: What are some
of the most important challenges facing the
electronics industry today?
STEVE MARTIN: I think the number one
challenge is globalization and the supply chain
complexity that results from it. I see more globalization and a resulting leakage in collaboration between everyone in the supply chain.
Companies are engaging in manufacturing
product and distributing product in different
Continued on Page 58
51
DistributionResource|GlobalPurchasing
Continued from Page 51
tors say they are thankful for the continued drive for more electronics, which
is helping to temper the situation. The
drive for smarter cars, medical equipment, appliances, and other consumer
devices is helping boost sales for many
companies, and so is factory automation
as the trend toward producing products
closer to home gains steam. It all adds
up to an electronics marketplace poised
for continued slow growth in the second
half of the year, say Knight and others.
“It’s looking to be a ver y slow
recovery,” says Julie Yuan, managing
director of California-based regional
electronics distributor Amidon. “It’s
certainly become more of a challenging marketplace.”
THE NEW NORMAL?
Although conditions are improving,
most economists predict continued
challenges in the manufacturing sector
and the overall economy for the remainder of 2013. In a late May economic
update, the Manufacturers Alliance for
Productivity and Innovation (MAPI)
predicted 3% growth in the manufacturing sector this year—a moderate pace,
but about a percentage point faster than
the general economy. Looking further
out, MAPI predicts 3.6% growth in the
sector in 2014, also about a percentage
point better than the general economy.
“The major constraint on consumers
in 2013 is that wages are not growing
much faster than the inflation rate and
spendable income is further reduced
by higher taxes,” MAPI chief economist
Daniel J. Meckstroth said, adding that
despite volatility and the struggle for
growth there are reasons for optimism
between now and the end of the year.
Rising home prices, pent-up demand
for consumer durables, and more balanced job growth between manufacturing, mining/construction and service industries are a few key reasons.
Despite that optimism, electronics
industry executives look to the remainder of 2013 with a cautious eye, and
52
Mouser Electronics is looking to expand its
Americas business with new growth in Central and South America, says Steve Newland,
vice president of sales for the Americas.
many are seeking growth by expanding their reach globally. Amidon is one
such company. Yuan says the distributor
will open an office in Hong Kong later
this year to capitalize on its growing
customer base in Asia. Similarly, large
catalog distributor Mouser Electronics
is looking to expand in the Americas. It
is in the process of opening a new office
in Mexico and is conducting a feasibility study to open one in Brazil. Expansion in the Americas follows Mouser’s
focused growth in Europe and Asia over
the last few years.
“I see automation really shaping [the industry],” says Steven Engineering’s Paul Burk.
“Customers are asking to automate processes that they never have before.”
“We’re starting to get more focused on
[Central and South America], [because]
we’ve seen some nice growth in Mexico
and Brazil,” explains vice president of
sales for the Americas Steve Newland,
pointing to Mouser’s mature business in
North America.
For many, mature North American
markets are complicated by the slow
recovery here at home. Looking at U.S.
manufacturing in particular, economic
activity had slowed as of early June compared to the beginning of the year.
The Institute for Supply Management’s Purchasing Manager’s Index
(PMI), which measures U.S. manufacturing activity, declined steadily from
February to May, when it hit its lowest
point in four years. May’s PMI registered
49, a nearly 2% slide compared to April
and signaling a decline in manufacturing activity.
A PMI above 50 indicates growth in
the sector. A PMI below 50 indicates
contraction. May’s reading was the first
contraction since November 2012 and
only the second since July 2009. Purchasing managers interviewed for the
survey indicated a flattening of demand
due to the sluggish local and global
economies.
REGIONALIZATION, AUTOMATION
KEY FACTORS
Rising costs are another key concern throughout the supply chain.
When it comes to transportation,
logistics, and labor, such increases are
signaling a new manufacturing trend
that bodes well for many distributors. Some North American executives point to a trend toward regional
manufacturing, also referred to as
on-shoring or re-shoring, as original
equipment manufacturers (OEMs)
seek to build their products closer to
where they will be consumed, especially large, heavy products that are
costly to ship around the world.
The situation is heightened as labor
costs increase in traditionally low-cost
regions such as Asia. Avnet Electronic
07.11.13 ELECTRONIC DESIGN
SIMPLIFIED.
SUPPLY CHAIN SOLUTIONS FROM AVNET
Your job is to develop and market innovative products; our job is to reduce your risk along the way.
Even leading edge products need a cost-effective and dependable supply chain to be a success. That’s
where Avnet can help. Around the world and around the clock, you can rely on our expertise at every
point in the supply chain – from supply chain assessment to solution design to global planning.
Avnet’s supply chain solutions save you inventory and infrastructure. That’s money in the bank, and
the edge you need to stay ahead of the competition. Simple, really.
What can we do for you? www.avnetexpress.com
Accelerating Your Success!™
1 800 332 8638 | www.avnetexpress.com |
@avnetdesignwire
DistributionResource|GlobalPurchasing
Marketing’s Ed Smith says the trend
is helping to increase manufacturing
activity in Mexico, as one example.
“I think what’s happening is, companies are saying, ‘I have certain products
that are worth building in Asia and certain products I need to build in America
[because] it’s not worth the logistics cost
to build them in Asia and ship them
back,’” says Smith, Avnet EM’s president
for the Americas, adding, “I don’t know
if I’d call it on-shoring or re-shoring, but
there is clearly an increase in manufacturing [in places such as] Mexico. Some
may view it as on-shoring or re-shoring.
I think there’s just less going out and
more being built in the regions where
it’s being consumed.”
Angelo Hrenczuk, director of sales,
thermal materials, for electronics manufacturer Laird Technologies, agrees
and says he sees more automotive and
appliance manufacturing returning to
Mexico as well.
“The chasing of low-cost centers is
going to have to end,” Hrenczuk says.
“Anything that’s big and expensive to
ship is coming back.”
Hrenczuk adds that an increase in
manufacturing in North and Central
America will also signal an increased
need for factory automation as manufacturers seek to become more competitive—another boon to many electronics
distributors’ business.
Northern California-based Steven
Engineering is already reaping the
rewards of that trend, says vice president
of marketing Paul Burk. The specialty
distributor of electronic components
and industrial automation equipment
is predicting double-digit growth this
year, with factory automation work
driving much of it. Burk points to medical/biotech, alternative energy (mainly
wind), and food and beverage processing equipment as key growth areas for
Steven Engineering.
“I see automation really shaping [the
industry],” says Burk. “Customers are
asking to automate processes that they
never have before.” 54
Capacitor Suppliers Meet
Growing Market Demands
The global capacitor market will grow 2.5% over the next
five years, driven by demand for more complex electronics
and innovative component solutions.
THE CAPACITOR MARKET is expected to
reach $20.2 billion in revenue by 2018,
according to a new report from industry
research firm Research and Markets.
The report predicts a 2.5% compound
annual growth rate for capacitors worldwide, driven by trends that continue to
affect the overall electronics industry,
including demand for compact, portable, and more complex electronic
devices and the accompanying need for
better, smaller, and more efficient component solutions.
The anticipated increase follows
a decline in the global capacitor market in 2012 due to slowing shipments
of PCs and televisions, weaker production capacities for wireless handsets, and
lower consumer and business spending on electronics, the report says. The
industry also suffered from the lingering
effects of natural disasters in 2011, particularly the Japanese tsunami early in
the year and severe flooding in Thailand
in the second half.
But the growing amount of electronic
content that’s employed in automotive
applications and the increased demand
for consumer electronics such as notebooks, ultrabooks, and smart phones
are key trends driving capacitor growth
over the long term. Component manufacturers are responding with enhanced
products and solutions that offer greater
efficiency and compact size to meet the
new demands. Distributors were promoting many of these new offerings
online this spring.
Kemet’s T488 Small Case Substrate Terminal MnO2, T527 Facedown Polymer, and
T529 Small Case Substrate Terminal Polymer
tantalum chip capacitors offer the highest
capacitance values available in both a standard 2012 and 3216 EIA case size with maximum package heights of as low as 1 mm.
ON THE MARKET
T488 Small Case Substrate Terminal
MnO2, the T527 Facedown Polymer,
and the T529 Small Case Substrate Terminal Polymer. These miniaturized solutions offer the highest capacitance values
(22 to 220 μF) available in both a standard 2012 (2 by 1.2 mm) and 3216 (3.2
by 1.6mm) Electronic Industries Association (EIA) case size with maximum
package heights of as low as 1 mm.
The capacitors also feature advances
in counter-electrode technology and
package design, resulting in the lowest
equivalent series resistance (ESR) values available in the industry for these
package sizes (see the figure). Typical
markets include computer/consumer
and telecommunications. Mouser Electronics (www.mouser.com) is among the
top distributors that are promoting the
new offerings.
Regional distributor Tonar Industries (www.tonar.com) is promoting
Kemet Corp. released a series of new
tantalum chip capacitors in May: the
Continued on Page 56
07.11.13 ELECTRONIC DESIGN
The Best in Automation & Control
Allied stocks all the top brands under one roof.
®
Play to Win!
Visit alliedelec.com/control through Aug. 31
for your chance to win some top prizes.
1.800.433.5700
© Allied Electronics, Inc 2013. ‘Allied Electronics’ and the Allied Electronics logo are trademarks of Allied Electronics, Inc.
An Electrocomponents Company.
DistributionResource|GlobalPurchasing
has issued guidelines that outline the
key aspects of compliance. He also said
companies should focus internally on
four key areas as they develop compliance strategies:
Few Companies Ready
For Conflict Minerals Regs
An IHS poll reveals that more than a third of electronics industry firms have made no plans to meet new requirements.
FEW U.S. COMPANIES
say they are prepared to meet new government regulations for reporting their use of so-called
conflict minerals, according to industry
analyst IHS. The regulations, which took
effect in August 2012 with initial reporting to begin in May 2014, are part of the
Securities and Exchange Commission’s
Dodd-Frank Wall Street Reform and
Consumer Protection Act of 2010.
IHS polled 134 electronics industry
managers this spring and found that just
7.5% said they were well prepared to
meet compliance rules. More than 35%
of respondents said they haven’t started
developing compliance plans, and the
remainder of respondents were somewhere in the middle.
The new conflict minerals requirements apply to publicly traded companies but will have far-reaching implications for the electronics supply chain.
Because manufacturers procure products and materials from all over the
world, it’s likely that one or more supplychain partners will require information
regarding the sourcing of conflict minerals, IHS explains.
Conflict minerals are raw materials mainly sourced in the Democratic
Republic of Congo, where their trade
fuels criminal networks and perpetuates
violence in the region. The materials—
tantalum, tungsten, gold, and tin—are
56
used in the electronics market in everything from cell phones to pacemakers.
IHS estimates that 15 cents worth of tantalum was in every smart phone shipped
when Dodd-Frank was originally signed
in 2010, for example. In 2012, this would
amount to $93 million worth of tantalum in smart phones, IHS says.
Newark element14’s Ken Manchen,
director of safety, health, and environmental affairs for the electronic components distributor, has said the rule will
affect distributors and suppliers as they
must report back to customers on the
presence and origin of conflict minerals
in the components they supply.
“This requirement will apply to both
private and foreign suppliers, even if not
regulated by the SEC,” Manchen said last
fall. “A manufacturer will not be able
to prepare a report without gathering
reports from their suppliers.”
Compliance is costly and time-consuming, but IHS notes industry efforts to
help support the process. IHS’s content
solution strategist Scott Wilson points to
the smelting industry in particular.
“Smelters are a good control point, and
this simplifies how far back in the supply
chain companies have to go,” Wilson told
an IHS webinar audience earlier in April.
Wilson also pointed to resources
such as the Organization for Economic
Cooperation and Development, which
• Management systems: Most material
requirements planning (MRP) and
enterprise resources planning (ERP)
systems can track materials, but may
not be programmed to do so. Companies need to determine if their systems need additional capabilities.
• Identify and assess risk: Companies
should prioritize the suppliers that are
most likely to use or source conflict
minerals.
• Respond to the risk: If suppliers don’t
share information, companies should
consider alternative sources.
• Audit smelters: An Electronics Industry Citizenship Coalition (EICC)
standard provides the necessary guidance and content. Continued from Page 54
new products from Cornell Dubilier.
The manufacturer expanded its line of
SPCX and SPSX surface-mount aluminum polymer capacitors, which are now
available with lower ESR and higher
capacitance values, making them topperforming filter capacitors for boardlevel power conversion applications.
Both types are supplied in a surfacemount molded package with the same
footprint as a D size tantalum capacitor
in a thin profile of 1.9 mm with voltage
values ranging from 2 to 6.3 V dc.
ATC’s ultra-broadband capacitors
are among Digi-Key’s (www.digikey.
com) newest passive component offerings. They offer reliable and repeatable
ultra-broadband performance from 16
kHz through more than 40 GHz. With
ultra-low insertion loss, flat frequency
response, and excellent return loss, the
550L series is ideal for dc blocking, coupling, bypassing, and feedback applications requiring ultra-broadband performance, the companies say. 07.11.13 ELECTRONIC DESIGN
We know you’re looking for outstanding value,
excellent customer service, proven market
intelligence, engineering expertise,
state-of-the-art distribution facilities, global
reach, on-time delivery and the most reliable
security of supply in the industry…
YOU GOT IT!
www.FutureElectronics.com
1-800-FUTURE-1
DistributionResource|GlobalPurchasing
Continued from Page 51
parts of the world. This creates many challenges in ensuring
that all standards of production and distribution are correctly
adhered to and that, ultimately, the traceability of that product back to the original manufacturer has not been obscured
in any way. Keeping a tight control on this process has been
something our industry is continuing to deal with.
Another challenge is the reduction in the product lifecycle
of more and more components, which ultimately leads to an
increasing rate of obsolescence in the supply chain. As manufacturers reduce the lifespan of many parts, and innovations in
technology are occurring at faster and faster rates, this results
in an increase in parts becoming obsolete and no longer produced by the manufacturer. The demand for this product,
however, may still be present, resulting in a severe mismatch
between supply and demand.
DR: Counterfeit components are another challenge. What
are the key issues buyers should consider when trying to
keep counterfeit components out of their supply chains?
MARTIN: This is a really wide topic of discussion. There is
a tremendous amount of things they need to do, but a few
come to mind first. They need to work with a trusted supplier
and have a continuous plan for vendor reduction. Also, they
<RXU
*OREDO
'LVWULEXWRU
RI
6SHFLDOW\
&KHPLFDOV
IRU
(OHFWURQLFV
$GKHVLYHV6HDODQWV&RDWLQJV
3RWWLQJ(QFDSVXODQWV7DSHV
DQG$SSOLFDWLRQ(TXLSPHQW
,62$6'LVWULEXWRU
ZZZHOOVZRUWKFRP
58
need to buy parts direct from the manufacturer or through
the franchised or authorized distributor. If they go outside
the authorized channel, there are a few telltale signs to look
for: how many years a company has been in business; the
size of the company in terms of revenue; and whether they
are members of any industry trade groups. We’re part of [the
Electronic Components Industry Association], for example.
In terms of warehousing, buyers need to know what the quality is, the handling, and the inspection process, both ingoing
and outgoing. Does the supplier have a counterfeit screening
or testing process in place? Are they ISO (International Organization for Standardization) certified? A lot of people say
they’re not certified but they are compliant. That should be a
telltale sign. A supplier really needs to be ISO certified to be
sure they’re on the up and up.
DR: Given the prevailing economic uncertainty worldwide,
what is your outlook for electronics distribution this year?
MARTIN: Even though there’s an ongoing recession in
Europe and slow growth in China, we are optimistic. Last
year was a building-block year. We expect right around 5%
growth in 2013.
Another growth area within the electronics distribution
industry that I predict we will continue to observe is an
increase in the growth of online purchases made by many
design and procurement engineers. People are beginning
to recognize the important role that e-commerce plays
within electronic component distribution, not just in facilitating online sales for parts, but also researching data on
parts as well as price point variables.
DR: What end markets or geographies look most promising to you?
MARTIN: In terms of excess, obsolete, and EOL (end of
life) components, we’re seeing that North America is very
promising. In addition, our e-commerce platform has seen
substantial growth in the last four months [in places such
as] Asia and South America. We see Brazil as a growing
market, and we expect growth in Mexico as well.
DR: Similarly, which end markets or geographies look
the most challenging?
MARTIN: Certainly Europe because of the economic difficulties there. But we’re also seeing some challenges in Asia.
There’s great opportunity in Asia, but I also think with the
counterfeit issue we’re seeing that as a challenging market
as well, particularly with the proliferation of counterfeit
parts in China. If you look at China in particular, there
are a lot of newer companies. Whether they are OEMs,
OCMs, distributors, or [businesses] in the secondary market, a lot of companies are popping up in China on a daily
basis, and there are no checks and balances on that. 07.11.13 ELECTRONIC DESIGN
NewProducts
AG
T
L
O
V
H
G
I
H
High-Performance Computing
Design Kit Optimizes Any SoC
Over
2500 Std. Models
Surface Mount and Thru-Hole
Power
Performance
D1
Delayed
optimized
flop
D
Q
Critical
path
logic
D
Q
Setup
optimized
flop
Clock signal
Master
latch
Slave
latch
Q1
Low Profile / Isolated
Up to 10,000 Volts Standard Regulated
and Programmable Models Available
Clock
D2
Delay optimized and setup optimized flops
manage critical logic paths
.-).%%%%.#&.*'-$'-(
Clock
Clock
E
Master
latch
Slave
latch
Q2
Multi-bit flops minimize loading, area, and leakage
The Synopsys HPC Design Kit comes with a range of flip-flop designs. Using flops designed for
an initial delay and then setup can lengthen the critical path time (left). Lower loading, smaller
size, and reduced leakage can be achieved using shared logic designs (right).
HOW DO you get the optimum systemon-chip (SoC) design? Tap the work
of other experts in the field. The Synopsys Designware HPC Design Kit delivers more than 125 design cells and
memories optimized for speed, space,
and power. It can be applied to SoC
designs that incorporate single-core
or multicore CPUs, GPUs, and DSPs. It
supports the latest TSMC 28-nm process today and is expected to extend
this support to other nodes soon.
The new package can be applied
to any design. It is different from ARM’s
Processor Optimization Package
(POP), which targets specific architectures like the ARM CPU. POP-style
solutions are fine when dealing with
the main CPU, but this latest kit works
with any CPU, GPU, or DSP core. It is
probably unnecessary and possibly
impossible to mix these approaches,
but the new HPC kit will be ideal for
custom designs.
The HPC kit includes a range of
flip-flop designs (see the figure). Each
optimizes a different set of criteria. It
would be great to have one design
that is the smallest, fastest, and most
power efficient, but normally there
are tradeoffs so the fastest is not the
most power efficient.
Critical-path designs can benefit by
using a flip-flop that optimizes the delay from input to output, which adds
more time to the critical path. Likewise,
a flip-flop that can extend the setup
time also adds time to the critical
path, resulting in the ability to incorporate more logic between the two different flip-flop implementations. Logically
they are the same at both ends, but
the timing details are different.
In some instances, speed or critical-path timing is less of an issue. In
this case, space and power efficiency
can be optimized. Register designs
that allow clock logic to be shared
help. Synopsys calls these register
designs multi-bit flip-flops.
Also, the range of cells is an advantage when mixing CPUs, GPUs, and
DSPs in an SoC. CPUs may have a
very high clock rate compared to the
GPUs and DSPs in the mix. Using areaoptimized two-port memory ensures
reduced area and power implementation for GPUs.
The selection of design cells still requires tradeoffs, but targeted designs
permit optimization for particular sets
of attributes.
SYNOPSYS
www.synopsys.com
BILL WONG
talog
ll ca
O’s fu
IC
See P
.p
www
ec
icoel
diately
imme
cs.c
t ro n i
om
High
Power
$.-).%.#&.*'-$'-(
"!-(.'$.-).%.--(
+# +..&* ..++
#,!,. -).-).*",.,,
).($,.'"-!-!,(
+&* &*+& .+"
143 Sparks Ave., Pelham, New York 10803
Call Toll Free 800-431-1064 • FAX 914-738-8225
E Mail: info@picoelectronics.com
GO TO ELECTRONICDESIGN.COM
NewProducts
Digital RF MEMS Capacitor Maximizes
Smart-Phone Antenna Efficiency
Antenna frequency
tuning performance
CAVENDISH KINETICS now offers a tunable RF MEMS capaci-
Antenna efficiency
tor that lets smart-phone antennas cover a wider range of
frequencies to ensure good efficiency.
Antennas can’t achieve maximum radiation efficiency
unless they’re precisely tuned to the frequency of operation. However, most smart phones using LTE and other
broadband modulation schemes operate in multiple
bands that can extend from as low as 700 MHz to over
2300 MHz. Multiple antennas partially solve the problem,
but they take up too much space in an already restricted
package—and add cost as well.
The solution is to tune the antenna to the band in use.
The simplest way to tune an antenna is to add a variable capacitor that can change the resonant frequency.
Electronically variable capacitors have been available for
years but have some disadvantages. For example, some
digitally variable capacitors use MOSFET switches to add
or subtract capacitors to or from a parallel connected
network of capacitors. The on resistance lowers the Q of
the overall capacitance and, therefore, the efficiency.
50%
Traditional antenna
performance
40%
30%
20%
Instantaneous
performance
10%
LTE
band
GSM
band
0%
600
700 800
900
Frequency (MHz)
1000
Using the Cavendish high-Q digitally variable MEMS capacitor, tuning an antenna in increments significantly improves its efficiency
compared to un-tuned antennas. The improved efficiency ensures
reliable connections and higher data rates.
Cavendish solves this problem by using hundreds of
tiny MEMS capacitors that are selected digitally in groups
to vary the capacitance. The device behaves like a bank
of 32 high-precision high-Q capacitors using a 1P32T
switch with zero insertion loss. The capacitor has a
5-to-1 capacitance range from maximum to minimum,
making wide-tuning frequency ranges possible. A serial
peripheral interface (SPI) takes the digital code to vary
the capacitance.
The figure illustrates the difference in performance
between a standard un-tuned antenna and a Cavendish MEMS tuned antenna. Operating from 700 to 1000
MHz, the standard antenna efficiency is less than 40%.
But tuning the antenna resonance at different parts of
the range greatly improves the efficiency.
This improved efficiency means more transmitted or
received signal to ensure not only a reliable connection
to the cell site but also higher data rates. In some cases,
a smaller antenna can be used or an extra antenna
can be eliminated.
In addition to antenna tuning, the Cavendish digitally
variable capacitor can be used to design variable
impedance matching circuits, power amplifier networks,
variable filters, and phase shifters and other networks
that must be optimized for the operating frequency. It is
available in a 2- by 2- by 0.4-mm chip scale package
(CSP). Its operating voltage is 1.8 V. Production samples
are available now.
CAVENDISH KINETICS
LOUIS E. FRENZEL
www.cavendish-kinetics.com
60
07.11.13
ELECTRONIC DESIGN
Fast Precision SAR ADCs Battle For Sockets
LINEAR TECHNOLOGY and Maxim
Integrated have announced highprecision successive approximation
register (SAR) architecture analog-todigital converters (ADCs) for the industrial and instrumentation markets.
Linear’s is faster with higher precision
while Maxim’s has a smaller footprint
and integrates its own voltage reference, saving more board space.
Linear’s 1-Msample LTC237820-1 boasts an internal clock and
accepts external references that
can range from 2.5 to 5.1 V. It offers
±2-ppm integral non-linearity (INL)
and –0.5 differential non-linearity
(DNL) (max), no missing codes at
20 bits, and a 104-dB signal-to-noise
ratio (SNR).
Also, the LTC2378-20-1’s data-sheet
guaranteed minimum signal-to-noise
plus distortion (SINAD) with a 5-V
reference and a 2-kHz frequency
is 101 dB, which calculates out to
a 16.48 effective number of bits
(ENOB). The ADC comes in a 16-lead
mini small-outline package (MSOP)
and a 4- by 3-mm dual flat no-lead
(DFN) package. In lots of 1000 units,
it costs $29.50 each. Linear offers less
expensive speed grades as well.
Maxim Integrated’s MAX11156 is
an 18-bit, 500-ksample/s, differential
input SAR ADC with a precision internal reference that allows for measuring a bipolar input voltage range
of ±5 V. With an external reference, it
can handle input ranges between
±3.05 V and ±5.19 V.
A patented charge-pump architecture permits direct sampling of
high-impedance sources. The ADC
also guarantees 18-bit no-missing
codes. The inputs are sampled with
a pseudo-differential on-chip trackand-hold that exhibits no pipeline
delay or latency.
The MAX11156’s SINAD is 91.5 dB
minimum for a 14.9 ENOB. Its SNR is
GO TO ELECTRONICDESIGN.COM
93 dB, based on the internal reference.
Standard packaging is a 12-pin, 3- by
3-mm thin DFN (TDFN). Unit pricing for
the MAX11156 with built-in voltage
reference is $16.90.
LINEAR TECHNOLOGY
www.linear.com
MAXIM INTEGRATED
www.maximintegrated.com
DON TUITE
61
DirectConnection
Texas Instruments is looking to fill engineering positions
working with semiconductor technologies. We are
recruiting @ various educational and experience levels:
Grass Valley, CA
• Analog Design Engineer - http://tinyurl.com/ol88wjg
Norcross, GA
• Analog Design Engineer - http://tinyurl.com/p8t35vr
San Diego, CA
• Applications Engineer - http://tinyurl.com/noob7jk
Stafford, TX
• Applications Manager - http://tinyurl.com/pfv2kpa
• Applications Engineer - http://tinyurl.com/nq8ocob
• Digital Design Engineer - http://tinyurl.com/q8gkw4m
• Electrical Design Engineer - http://tinyurl.com/qb7r4sp
• Product Engineer - http://tinyurl.com/np79qzk
• Software Design Engineer - http://tinyurl.com/owbqncx
• Software Engineer - http://tinyurl.com/q7xzjag
IR’s New 1200V IGBTs Deliver Higher Power Density and
Increased Efficiency for Motor Drive and UPS Applications
IR has introduced a family of rugged 1200V ultra-fast
insulated-gate bipolar transistors (IGBTs) optimized for
industrial motor drive and UPS systems.
The new devices leverage IR’s field stop trench ultrathin wafer technology that delivers lower conduction and
switching losses. Co-packaged with a soft recovery low
Qrr diode and featuring a 10us minimum short circuit
time rating, the devices are optimized for rugged industrial
applications.
The packaged devices cover a broad current range from 10 – 50A. Other key performance
benefits include Tjmax of 150°C, positive VCE(on) temperature coefficient for easy
paralleling and low VCE(on) to reduce power dissipation and achieve higher power density.
Die products are also available.
More information is available on the International Rectifier website at
http://www.irf.com/whats-new/nr130606.html
For more information, contact Sian Cummins,
scummin1@irf.com, 310-252-7148.
• Systems Engineer - http://tinyurl.com/qj2ep64
• Systems Manager - http://tinyurl.com/owyl585
The NEW
Woodland Hills, CA
• Tech Sales Rep/Eng - http://tinyurl.com/ppht2zc
Some positions may require some travel. Please
The place where engineering videos gather
send in resumes using the individual URL code for the
position listed. EOE.
Catalog Literature
Review
Engineering Channels
HIGH PERFORMANCE SOCKETS
& ADAPTERS
,QFOXGHV *+] EDQGZLGWK
VRFNHWV IRU %*$ DQG 4)1
RQO\ VOLJKWO\ ODUJHU WKDQ ,&
ZLWK LQWHJUDO KHDWVLQN IRU
PHGLXP SRZHU DQG RSWLRQDO
KHDWVLQNLQJ WR RYHU ZDWWV 8S WR N LQVHUWLRQV
$GDSWHUV IRU SURWRW\SH WHVW
SDFNDJH FRQYHUVLRQ DQG
PRUH 4XLFN7XUQ YROXPH
DGDSWHUVDUHRXUVSHFLDOW\
Ironwood Electronics
1335 Eagandale Court
Eagan, MN 55121
T: 800-404-0204
F: 952-229-8201
www.ironwoodelectronics.com
62
Recent Videos
Trade Show Coverage
THE SAME GREAT VIDEO CONTENT, IN AN EASY-TO-BROWSE FORMAT
In the past, we’ve been told by our design
engineers that Engineering TV is their favorite
place to “geek out”. Well get your computers
ready, because Engineering TV is all new and
completely redesigned!
Browsing the most compelling videos is
easier than ever. Our new engineering channels
let you explore the technologies that are of
interest to you – everything from Aerospace to
Embedded to Robotics. Plus, we film at all the
key electronics and design engineering trade
shows, so even if you can’t attend, you can
find out everything that is happening at these
important events.
CHECK US OUT TODAY:
www.engineeringtv.com
ELECTRONIC DESIGN
DirectConnection
Otter Computer, Inc.
Spring Pin Socket for Kelvin Resistance
Measurement
Power management devices need to
be measured precisely for resistance
values. Kelvin technique allows
precise resistance measurement of
devices using BGA or QFN socket
over a wide temperature range with
Ironwood Electronics’ Kelvin socket
technology. BGA and QFN GHz
bandwidth Sockets for pitches down to 0.4 mm pitch and
bandwidth to 40 GHz, with industry’s smallest footprint
simplifies development cycle.
Ironwood Electronics, Inc.
1335 Eagandale Court
Eagan, MN 55121
T: 800-404-0204 • F: 952-229-8201
Email: info@ironwoodelectronics.com
Web Site: www.ironwoodelectronics.com
FCI
Ad
Page
Ad
Page
Ad
Page
ACCES I/O Products Inc. ............................. 47
Ellsworth Adhesives ...................................... 58
National Instruments ..................................... 12
Agilent Technologies .................................... 11
ExpressPCB .................................................. 32
Newark ............................................................ 3
Agilent Technologies .................................... 17
FCI ................................................................ 25
Pico Electronics Inc. ..................................... 33
Agilent Technologies .................................... 31
Front Panel Express ...................................... 40
Pico Electronics Inc. ..................................... 59
Allied Electronics.......................................... 55
Future Electronics ......................................... 57
Rohde & Schwarz ...................................... IBC
ARM ............................................................. 10
Heilind Electronics........................................ 60
Shenzhen Fastprint Circuit Tech Co. Ltd. ..... 61
Avago Technologies ................................. 32a/b
Interconnect Systems Inc. ............................. 35
Stanford Research Systems ......................26-27
Avnet ............................................................... 9
International Rectifier ..................................... 4
Avnet ............................................................. 53
Ironwood Electronics .................................... 45
Beta-Layout................................................... 40
Linear Technology ................................... 48a/b
Coilcraft .......................................................... 1
Linear Technology .......................................BC
CTS Electronic Components ........................ 46
Linx Technologies ......................................... 34
CUI Inc.......................................................... 18
Maxim Integrated .......................................... 23
Digi-Key....................................................... FC
Mouser Electronics ......................................... 6
Digi-Key......................................................IFC
Mouser Electronics ....................................... 50
ELECTRONIC DESIGN*272:::(/(&7521,&'(6,*1&20
Supertex Inc. ................................................. 42
Taiyo Yuden ................................................... 15
TDK-Lambda ................................................ 19
Zilog .............................................................. 39
For more information on products or services visit our
website www.electronicdesign.com, menu item Reader
Service. The advertisers index is prepared as an extra
service. Electronic Design does not assume any liability for
omissions or errors.
63
LabBench
BILL WONG | Embedded/Systems/Software Editor
bill.wong@penton.com
Are You Writing
Safe And Secure Software?
S
ecure Boot, iPhones, mulch, safety, and
Ada—I promise to tie them all together,
since they’re all related. Sort of.
As a programmer, I try to write code that’s not
only functional but also safe, secure, and reliable.
I work at learning the best techniques, using the
best compiler tools, and taking advantage of the
latest technology. Great. Everyone knows that C
dominates the embedded space. But what if some
of those things I learn or use aren’t quite right?
The new USB standard also provides a highcurrent feature, but it does not use the data link.
Instead, it uses some creative analog signalling.
READ ABOUT ADA 2012, TO BE SAFE
This brings me to Ada, which is designed
from the ground up for safety and security. Its
features also turn out to make bug-free programming much easier. But from a programming perspective, I’m still a neophyte. I have
Georgia Tech students atbeen using it more and learning quite a bit
tacked a poor, helpless iPhone
about it, but I defer to the experts in trying to
using a string overflow bug in
... BUT EVERYONE USES MULCH
convince programmers about its advantages.
So now the mulch. You’ve seen it everywhere the phone’s USB charger DRM
Safe and Secure Software - An Invitation to
support designed to restrict
on landscaped lawns. There are giant mounds
Ada 2012, a free e-book from AdaCore availchargers to those licensed by
around every tree. It looks great, if you like that
able at www.adacore.com/safe-secure-booklet,
Apple.
kind of thing. It makes mowing, especially with
highlights programming issues that arise with
large riding lawnmowers, so easy.
any programming language. Of course, it explains why Ada
Unfortunately, piling bark mulch around most anything,
2012 is a better solution often because it does things that other
especially trees, isn’t a great idea. Compost is good around
languages, including C++ and Java, do not address.
bushes, but trees don’t like anything stacked up on their bark.
For example, Ada’s type system puts most others to shame.
There are other good uses for bark mulch, though.
Its pointer support provides finer-grain control, allowing the
C and C++ aren’t quite as bad as mulch around trees, but
compiler to catch more errors. Ada’s storage management is
using them just because everyone else does can be an issue
something that most languages relegate to libraries, but there
because they’re prone to bugs. Dangling pointers and buffer
are reasons for incorporating it into the compiler.
overflows are common C bugs.
Ada 2012 also brings contracts out of the SPARK comment
realm
and into the language itself. This is reason enough to take
... AND NOW THE IPHONE
another
look at Ada, but there is plenty more. The book shows
Researchers at my alma mater, Georgia Tech, attacked
the
problems,
how they are addressed with other languages, and
an iPhone using a smart USB charger (see the figure). Apple
why
Ada
was
designed
to address these issues.
decided to use digital rights management (DRM) to restrict
The
book
offers
a
flavor
of Ada, but it is designed for any prolicensing of high-current USB chargers. Essentially, the charger
grammer.
It
addresses
Ada
basics like built-in range checking
handshakes with the phone using the USB data channel before
that
probably
would
have
caught
the iPhone charger problem.
it raises the amount of power delivered.
Hopefully,
the
book
will
get
you
thinking.
The charger sends a key in its handshake packet. The
Still, C and C++ won’t be abandoned en masse. Tools like
researchers caused a buffer overflow. This form of injection
MISRA
C can help. I recommend looking into static and
attack is common in jailbreaking smart phones and other devicdynamic
analysis tools because they do make a difference.
es. They take advantage of a software bug.
The number of buffer overflow attacks on everything from
smart phones to servers is rather extensive, though. One might
Follow Bill Wong on Twitter at #AltEmbedded.
think that avoiding these kinds of problems would improve the
Become a fan at facebook.com/ElectronicDesign.
safety and security of quite a few systems.
64
07.11.13 ELECTRONIC DESIGN
'XGTYKUJGFHQTC
DGVVGTDGPEJUEQRG!
6JGPGY46/6WTPQP/GCUWTG
'CU[JCPFNKPIHCUVCPFTGNKCDNGTGUWNVUsGZCEVN[YJCVWUGTUGZRGEV
HTQOCDGPEJQUEKNNQUEQRG4QJFG5EJYCT\QRGPUVJGFQQTVQ
CPGYYQTNF9QTMYKVJVYQUETGGPUQPQPGFKURNC[#EEGUUCNN
HWPEVKQPUSWKEMN[#PCN[\GOGCUWTGOGPVTGUWNVUYJKNGQVJGTUCTG
UVKNNDQQVKPIWR5GGUKIPCNUYJGTGQVJGTULWUVUJQYPQKUG
6JCVoUVJG46/
'XGTYKUJGFVJGTGYCUCPGCUKGTYC[!'XGTYKUJGFHQTOQTG
TGNKCDNGTGUWNVU!'XGTYKUJGF[QWEQWNFFQ[QWTLQDHCUVGT!
6JGPVCMGCNQQM
YYYUEQRGQHVJGCTVEQOCFTVOXKFGQ
20V, 2.5A SWITCHER+
VIN
12V
BOOST
PVIN
47μF
VOUT
0.8V @ 2.5A
1.5μH
SW
INTVCC
47μF
FB
TTHERMAL LIMIT
TMONITOR
PGND
IMONITOR
FSW = 2MHz
Programmable Input & Output Current Limits and
Monitoring with Die Temp Sensing
®
™
The LTC 3626 SWITCHER+ is the first member of a new switcher family featuring programmable input and output
current limits, as well as on-chip die temperature monitoring with programmable thermal shutdown. These features
provide enhanced system level protection, control and real time status readings. Plus, its unique controlled on-time
architecture is ideal for high step-down ratios where high switching frequencies and fast transient response are essential.
The LTC3626 combines up to 3MHz switching with a small, compact solution footprint.
Features
sWide Output Voltage Range:
0.6V to 97%VIN
s95% Efficient
sUp to 2.5A Output Current
- Average Input and Output Current
Monitoring (IMONIN, IMONOUT)
- Programmable Average Input/Output
Current Limit
- Die Temp Monitor and Programmable
Limit (TSET)
Measured Output Current Error (%)
s2.6V to 20V Input Voltage
Output Current Monitor Error
5
Info & Free Samples
www.linear.com/product/LTC3626
4
1-800-4-LINEAR
3
2
1
0
–1
–2
–3
TA = 85°C
TA = 25°C
TA = –40°C
–4
–5
1
1.25
1.5
1.75
2
Output Current (A)
2.25
2.5
cds.linear.com/docs/Design%20Note/dn511.pdf
, LT, LTC, LTM, Linear Technology and the Linear logo are
registered trademarks and Switcher+ is a trademark of Linear
Technology Corporation. All other trademarks are the property
of their respective owners.